Dragoon952
The Winter Moth
Double post *shakes fist at Serebii server*
-
Hi all. We have had reports of member's signatures being edited to include malicious content. You can rest assured this wasn't done by staff and we can find no indication that the forums themselves have been compromised.
However, remember to keep your passwords secure. If you use similar logins on multiple sites, people and even bots may be able to access your account.
We always recommend using unique passwords and enable two-factor authentication if possible. Make sure you are secure. -
Be sure to join the discussion on our discord at: Discord.gg/serebii
A Possible Alternative to Smogon Tiers
- Thread starter Dragoon952
- Start date
Dragoon952
The Winter Moth
UPDATE
Finished the "tweaks" and taking the 75%/50% stealth rock modifiers into account. I updated the first post with the brand new numbers. I made two lists: one alphabetical and one by value.
Also, some people were curious what the values would look like in the 100's instead of the 10's, so I included both of those numbers. I'm still partial to the simpler values in the 10's, but now you can see what it looks like both ways.
EDIT:
Is there any easy way to clean up those numbers so they appear in columns?
Finished the "tweaks" and taking the 75%/50% stealth rock modifiers into account. I updated the first post with the brand new numbers. I made two lists: one alphabetical and one by value.
Also, some people were curious what the values would look like in the 100's instead of the 10's, so I included both of those numbers. I'm still partial to the simpler values in the 10's, but now you can see what it looks like both ways.
EDIT:
Is there any easy way to clean up those numbers so they appear in columns?
Steelix211
New Member
By the way guys, I've just finished crunching all of the numbers and got around 70-78 as a possible team value total. Does that seem a bit high/low to anyone? I left out Pokemon such as the Rotom Formes, Miltank and Slowking as they don't have their values yet. Would Slowking have the same value as Slowbro, or would Nasty Plot and its stat arrangement make a significant difference?
Prince Amrod
Dragon Tamer
i think that due to slowkings stats nd access to nasty plot, it shud be placed higher than slowbro, it makes enough of a difference i think to make it a greater valued pokemon over its weaker cousin
Dragoon952
The Winter Moth
70 might be good. That's an average of between 11 and 12 per pokemon on a 6 pokemon team. I'm sure you could drop it down to 65 which would be closer to just 11 per pokemon. People could try 70 in a few test runs and see what diversity comes back, and if it just looks like an OU battle you could drop it to 65.
Dragoon952
The Winter Moth
Ok, quick update. Someone pointed out to me that the Rotom forms are off and I missed their different base stats. So, here are the actual Rotom numbers:
Rotom 11 109
Rotom (Fan) 12 124
Rotom (Frost) 13 125
Rotom (Heat) 12 122
Rotom (Mow) 12 122
Rotom (Wash) 12 124
I updated the first post with that info.
Rotom 11 109
Rotom (Fan) 12 124
Rotom (Frost) 13 125
Rotom (Heat) 12 122
Rotom (Mow) 12 122
Rotom (Wash) 12 124
I updated the first post with that info.
RhythmScript
New Member
This thread has seriously been going strong since last year. Wow.
So yes, I actually created an account just to post in this thread. Recently I've been looking at getting into competitive battling and became well-versed in Smogon and their rules and systems and I remember it always bugged the crap out of me how their limitations worked. Pokemon got declared "broken" when they were too powerful, so they were banned, with an absolute lack of the foresight necessary to realize that the NEXT most powerful poke would take its place. I remember thinking "rather than just banning something that becomes too powerful, why don't you mark each pokemon's power on some sort of gradient and use a limit system to cripple teams with over-powered pokes by forcing them to pair them with under-powered pokes?"
I remember trying to figure out how this would work. I considered base stat totals, but realized that it didn't take typing into account. Considered using a max-per-team of each individual stat, but realized it wouldn't account for pokemon whose particular high-stat combination (speed and attack, for example) made them so deadly.
I've been working on this system, trying to figure out a basis for it, for the last two weeks. Yesterday, it occured to me that it might make sense to just give the pokemon a standard numerical rank based on their "usefulness" (arbitrated by players of the poke-verse far more skilled than I) and put a cap on the team's score limit. That way overpowered pokes would lead to underpowered partners, or perhaps, even teams with fewer than six pokemon.
Then, in a burst of random Googling (I honestly have no idea how I found this thread. I think I was looking up the viabiltiy of Draco Meteor on Jirachi) I stumbled upon this thread.
It kind of hurts to see my idea so thoroughy beaten to the punch by Dragoon and his close-knit cadre of contributors, but on the other hand, I'm glad to see that it's gotten through the hectic stages of early brokenness and come out stronger; if I had been working on this system I'd have abandoned it and started a game of Mario Galaxy a LONG time ago.
I've read through every post in this thread, starting on page one, and while I got lost in some of the Pythagorean math back on page 3 or 4, I'm starting to really appreciate the elegance of this scoring system, not only in its commitment to avoiding "special circumstances" for pokemon whenever possible, but also its attempts to develop a system that, once tweaked properly, should be able to comfortably account for newly introduced pokes, moves, abilities, and pretty much anything else the game throws in in future installments (barring a new battle mechanic entirely).
That said; there are a couple of concerns that I'd like to raise regarding more recent developments of the formula (though I should admit to not having downloaded the Excel files and had a proper look at the numbers).
ONE: Items - Yes, I'm aware we've covered this, particularly the reality that an item tends to land evenly in the point spread if all pokemon can use it. My suggestion is, instead, to make an item ITSELF cost points irrespective of who it's equipped to. Noble a pursit as it is to want to balance out the metagame by preventing the same cluster of pokes from being used, I don't feel I'm alone in being sick to death of seeing leftovers on every pokemon who doesn't have a choice item or Life Orb equipped.
I'm aware of the potential for cheating because an item can't be seen right at the outset. But right now there are few enough PVS battlers, most of whom want to see it succeed, that I don't think the honor system would fail very hard at this stage. And, assuming this system is picked up by Shoddy Battlers, it shouldn't be too hard to have Shoddy figure in the math itself, locking teams to their ceiling without having players do the math. Which brings me to my second point...
TWO: Get Shoddy on your side - I'm not sure if you're aware, but Shoddy Battle two is currently in development and is at a stage where implementing a system like this is certainly viable, especially if the math is already precalculated by someone other than the Shoddy programmers. If you could get someone on the inside at Shoddy to take an interest in this system, it would not only help the system gain curious newcomers, but it would make the math a lot easier for everyone involved. Attaching a single value field to each pokemon from a table doesn't seem like a lot of work compared to some of the other monotony faced by the Shoddy dev team, and if they find the system interesting I don't think they'd be unwilling.
The only reason I bring it up is I'm not sure anyone here has really broached the subject to them; if you make this visible to the people with their finger on the power button of the battling community it would go a long way toward getting the system publicized.
THREE: The Salamence issue - This is sort of unrelated, but it's topical so I'll go ahead with it. Salamence is probably going to be banned from the Smogon OU tier, blah blah blah that's not really important in and of itself. But what is important is the REASON; because of it's relatively high SpAtk AND Atk stats, a Salamence poses a threat not because it does one thing exceptionally well, but because it does two things very well. Of course, if you KNOW how a Salamence is going to play, like you would with, say, a Blissey, you can react accordingly. People are frustrated because a Salamence can easily play several different ways, all of them equally devastating, and just in figuring out which method is in use you'll probably lose a poke and a half. This was at the heart of what made Garchomp so difficult as well; it wasn't that he did one thing super-well, it was that figuring out which Chomp was being played against you was too costly a discovery.
I bring this up because I've noticed this kind of consideration being given to things outside of the base-stat realm; type coverage in the movepool, for example, is a concern because you don't know at the outset which moves a poke is running, so having access to more moves means more potential danger, even if any one combination isn't more lethal than on a different poke.
But is this consideration being given to pokemon whose STATS allow them to play a variety of ways? pokemon are typically EV'd to accentuate their base stats, because (barring certain circumstances) it doesn't make sense to boost an already low stat as it's a waste of EVs. But some pokemon have base stats which are clustered together enough that you can choose to play them in a number of ways, all equally successfully, and this makes them particularly dangerous.
A good example is Salamence, as I brought up before; Blissey is rated 228 by your numbering, while Mence is only 162. This may make sense for any particular set that Mence is running. But because it's impossible to tell whether you're up against, say, a DDMence or a MixMence until you've lost a poke, while Blissey is always going to be performing essentially the same role, Mence is a much bigger threat and more valuable poke. There's a reason he's up for Uber consideration and Blissey isn't.
Unfortunately, I don't know how to quantify this difference, but I think you have to consider the additional danger posed by the unpredictability of a pokemon in the metagame when figuring in its overall cost to a team. Maybe a system whereby any base stat which is greater than a certain value (80? 100? 110?) gets weighted, so you have more points being given to those pokes with more than one stat which is surprisingly high?
Anyway, /walloftext. I just wanted to say I'm super-excited about the system and, as I finish putting together my battling team, am very interested to see where it falls in the point spread and which pokemon I'll have to drop in order to test this system.
EDIT: Also, out of pure curiosity, how does Hidden Power factor into your math? I notice that you give consideration to the "most powerful move" of each type, isn't that what you said? Hidden Power is a strange issue because it can offer a BP 70 move of any type to essentially ANY pokemon (at least, any which can learn TMs, which IIRC is all but like 15). But, it can only be used for one type, meaning you can't get, say, Ice coverage AND Ground coverage on a poke which learns neither of those move types, but you can get one or the other (which is distinct from how the other learnset moves operate). It might seem like it'd be easiest just to ignore HP as a move from the calculations but some pokemon rely on it to round out their sets, and when a pokemon like Magnezone is granted access to a base 70 power Grass, Fire, or Ice type move it drastically changes its viability in the game, moreso than HP would for other pokes which already HAVE good type coverage. I guess my point being that there are some pokes which are probably undervalued by your math (assuming HP is disregarded) because it seems like they're lacking in type coverage, but a well-selected Hidden Power will more than make up for that deficiency if their movepool only has a hard time covering one or two types (Ice HP on pokemon who otherwise can't deal with dragon types AT ALL being a prime example).
Just some food for thought.
So yes, I actually created an account just to post in this thread. Recently I've been looking at getting into competitive battling and became well-versed in Smogon and their rules and systems and I remember it always bugged the crap out of me how their limitations worked. Pokemon got declared "broken" when they were too powerful, so they were banned, with an absolute lack of the foresight necessary to realize that the NEXT most powerful poke would take its place. I remember thinking "rather than just banning something that becomes too powerful, why don't you mark each pokemon's power on some sort of gradient and use a limit system to cripple teams with over-powered pokes by forcing them to pair them with under-powered pokes?"
I remember trying to figure out how this would work. I considered base stat totals, but realized that it didn't take typing into account. Considered using a max-per-team of each individual stat, but realized it wouldn't account for pokemon whose particular high-stat combination (speed and attack, for example) made them so deadly.
I've been working on this system, trying to figure out a basis for it, for the last two weeks. Yesterday, it occured to me that it might make sense to just give the pokemon a standard numerical rank based on their "usefulness" (arbitrated by players of the poke-verse far more skilled than I) and put a cap on the team's score limit. That way overpowered pokes would lead to underpowered partners, or perhaps, even teams with fewer than six pokemon.
Then, in a burst of random Googling (I honestly have no idea how I found this thread. I think I was looking up the viabiltiy of Draco Meteor on Jirachi) I stumbled upon this thread.
It kind of hurts to see my idea so thoroughy beaten to the punch by Dragoon and his close-knit cadre of contributors, but on the other hand, I'm glad to see that it's gotten through the hectic stages of early brokenness and come out stronger; if I had been working on this system I'd have abandoned it and started a game of Mario Galaxy a LONG time ago.
I've read through every post in this thread, starting on page one, and while I got lost in some of the Pythagorean math back on page 3 or 4, I'm starting to really appreciate the elegance of this scoring system, not only in its commitment to avoiding "special circumstances" for pokemon whenever possible, but also its attempts to develop a system that, once tweaked properly, should be able to comfortably account for newly introduced pokes, moves, abilities, and pretty much anything else the game throws in in future installments (barring a new battle mechanic entirely).
That said; there are a couple of concerns that I'd like to raise regarding more recent developments of the formula (though I should admit to not having downloaded the Excel files and had a proper look at the numbers).
ONE: Items - Yes, I'm aware we've covered this, particularly the reality that an item tends to land evenly in the point spread if all pokemon can use it. My suggestion is, instead, to make an item ITSELF cost points irrespective of who it's equipped to. Noble a pursit as it is to want to balance out the metagame by preventing the same cluster of pokes from being used, I don't feel I'm alone in being sick to death of seeing leftovers on every pokemon who doesn't have a choice item or Life Orb equipped.
I'm aware of the potential for cheating because an item can't be seen right at the outset. But right now there are few enough PVS battlers, most of whom want to see it succeed, that I don't think the honor system would fail very hard at this stage. And, assuming this system is picked up by Shoddy Battlers, it shouldn't be too hard to have Shoddy figure in the math itself, locking teams to their ceiling without having players do the math. Which brings me to my second point...
TWO: Get Shoddy on your side - I'm not sure if you're aware, but Shoddy Battle two is currently in development and is at a stage where implementing a system like this is certainly viable, especially if the math is already precalculated by someone other than the Shoddy programmers. If you could get someone on the inside at Shoddy to take an interest in this system, it would not only help the system gain curious newcomers, but it would make the math a lot easier for everyone involved. Attaching a single value field to each pokemon from a table doesn't seem like a lot of work compared to some of the other monotony faced by the Shoddy dev team, and if they find the system interesting I don't think they'd be unwilling.
The only reason I bring it up is I'm not sure anyone here has really broached the subject to them; if you make this visible to the people with their finger on the power button of the battling community it would go a long way toward getting the system publicized.
THREE: The Salamence issue - This is sort of unrelated, but it's topical so I'll go ahead with it. Salamence is probably going to be banned from the Smogon OU tier, blah blah blah that's not really important in and of itself. But what is important is the REASON; because of it's relatively high SpAtk AND Atk stats, a Salamence poses a threat not because it does one thing exceptionally well, but because it does two things very well. Of course, if you KNOW how a Salamence is going to play, like you would with, say, a Blissey, you can react accordingly. People are frustrated because a Salamence can easily play several different ways, all of them equally devastating, and just in figuring out which method is in use you'll probably lose a poke and a half. This was at the heart of what made Garchomp so difficult as well; it wasn't that he did one thing super-well, it was that figuring out which Chomp was being played against you was too costly a discovery.
I bring this up because I've noticed this kind of consideration being given to things outside of the base-stat realm; type coverage in the movepool, for example, is a concern because you don't know at the outset which moves a poke is running, so having access to more moves means more potential danger, even if any one combination isn't more lethal than on a different poke.
But is this consideration being given to pokemon whose STATS allow them to play a variety of ways? pokemon are typically EV'd to accentuate their base stats, because (barring certain circumstances) it doesn't make sense to boost an already low stat as it's a waste of EVs. But some pokemon have base stats which are clustered together enough that you can choose to play them in a number of ways, all equally successfully, and this makes them particularly dangerous.
A good example is Salamence, as I brought up before; Blissey is rated 228 by your numbering, while Mence is only 162. This may make sense for any particular set that Mence is running. But because it's impossible to tell whether you're up against, say, a DDMence or a MixMence until you've lost a poke, while Blissey is always going to be performing essentially the same role, Mence is a much bigger threat and more valuable poke. There's a reason he's up for Uber consideration and Blissey isn't.
Unfortunately, I don't know how to quantify this difference, but I think you have to consider the additional danger posed by the unpredictability of a pokemon in the metagame when figuring in its overall cost to a team. Maybe a system whereby any base stat which is greater than a certain value (80? 100? 110?) gets weighted, so you have more points being given to those pokes with more than one stat which is surprisingly high?
Anyway, /walloftext. I just wanted to say I'm super-excited about the system and, as I finish putting together my battling team, am very interested to see where it falls in the point spread and which pokemon I'll have to drop in order to test this system.
EDIT: Also, out of pure curiosity, how does Hidden Power factor into your math? I notice that you give consideration to the "most powerful move" of each type, isn't that what you said? Hidden Power is a strange issue because it can offer a BP 70 move of any type to essentially ANY pokemon (at least, any which can learn TMs, which IIRC is all but like 15). But, it can only be used for one type, meaning you can't get, say, Ice coverage AND Ground coverage on a poke which learns neither of those move types, but you can get one or the other (which is distinct from how the other learnset moves operate). It might seem like it'd be easiest just to ignore HP as a move from the calculations but some pokemon rely on it to round out their sets, and when a pokemon like Magnezone is granted access to a base 70 power Grass, Fire, or Ice type move it drastically changes its viability in the game, moreso than HP would for other pokes which already HAVE good type coverage. I guess my point being that there are some pokes which are probably undervalued by your math (assuming HP is disregarded) because it seems like they're lacking in type coverage, but a well-selected Hidden Power will more than make up for that deficiency if their movepool only has a hard time covering one or two types (Ice HP on pokemon who otherwise can't deal with dragon types AT ALL being a prime example).
Just some food for thought.
TheFightingPikachu
Smashing!
Joined just to post here? That, is actually pretty cool.
People can actually find this thread by searching the internet?! Okay, I'm being a bit sarcastic, but I truly would not have guessed that. Amazing!
Hmmm....True, the honor system wouldn't fail very hard at this stage. I'm not trying to be antagonistic when I say that Leftovers on every Pokemon isn't that big a deal. I actually have heard Leftovers recommended as the default item for any Pokemon without a specific item-based strategy. I can't see that as something worth restricting.
I hope I'm not being either too cynical for suggesting that we might be laughed out of town by Smogon, nor being too sentimental for wanting this system to gain its popularity apart from Smogon's influence.
Regarding Salamence, yes, I'd say you are correct. Salamence does appear to have too little value at present. I'm not really sure what should be done about it.
Dragoon952
The Winter Moth
First of all, I'm very glad you found this thread. You seem to think a lot like I do about things 
I'll let you decide if that's good or not.
The more contribution, the better this system will be. So I'm gladd to hear your thoughts.
I guess part of the problem is that some of us are thinking in terms of Wifi. There's really no easy way to implement such a system over wifi, and you don't want people to have to tip their hand in advance. I think assigning values to items is the best way to go, it's just a question of implementation. Not to mention that a set number increase will more drastically change lower valued pokemon and barely dent upper valued ones (unless it is a percentage change of the base value). An item valued at 3 would boost a pokemon with a value of 6 by 50%, but would boost a pokemon with a value of 15 only 20% (if my math is right). Seems like it should be the other way around.
Another option, albeit a drastic one, is to go the Battle Tower route and limit a battle by not allowing more than one of any item. So, you'd have to choose where the leftovers go. Still chance for cheating, but less in my opinion. Easier to find two of the same item than trying to guess and do the math in your head.
I think items can be revisited later, but good thoughts all around.
I think I agree with TFP, even though it does seem counterproductive. I've lurked over at smogon, and I just get the general sense that they won't be very accpeting of the ideas presented here.
To be honest, I'm one guy that is using free time to do this for fun. I hit quite a bit of criticism right off the bat, and most of it was along the lines of "it can't be done and don't even try" rather than "I don't think this part of the math is accurate." I think most of the "constructive criticism" would fall under the first category if this was presented over at smogon. Particularly because everyone has their own opinion on which pokemon rocks, so to speak.
That aside, I do envision publicizing this more in the future for sure. I just think people go crazy when this thing is still in its infancy. They want it perfect right out of the gate or not at all, which doesn't make sense to me. So, I guess I want to make sure it is "presentable" before it goes further.
Although I wouldn't be opposed to someone bringing it up over there and directing them here
Now it's your turn to reach in my brain before I get there
I saw the Salamence discussion over at smogon. I agree with your analysis, and lots of people over there seem to as well in regards to the fact that there will always be a "most powerful" pokemon and it will be never ending if you keep kicking them out.
Salamence WAS 22 (216)...before the Stealth Rock modifier I added in to the mix. There was a little discussion about trying to make the modifier less drastic, which might help. Need to look at that more.
Still, I had almost the exact same thought about base stats where there should be a bonus for super high stats. I had an idea of maybe adding a "Total Base Stat" bonus. It would be a general way of giving a bonus for pokemon that would have a stat spread where several stats are high.
So, I just picked a number and went from there. I figured 100 or higher in multiple stats would pick out pokemon like you are describing, so spread across all six stats that would give a base stat total of 600 (averaging it out to account for stats above and below that line on a pokemon). I took a look at the list at Serebii and that looks like a perfect number. It would give a bonus to the following pokemon and no others:
Arceus, Mewtwo, Lugia, Ho-oh, Rayquaza, Dialga, Palkia, Giratina, Slaking, Kyogre, Groudon, Regigigas, Dragonite, Mew, Tyranitar, Celebi, Salamence, Metagross, Latias, Latios, Jirachi, Deoxys, Garchomp, Heatran, Cressalia, Manaphy, Darkrai, and Shaymin.
Not only would that further boost Ubers (which most of those are), but it will give a boost to some of the other "borderline" type pokemon like Salamence. There are a few odd ones like Slaking and Regigigas, but they shouldn't be really affected too much because their abilities will severely hamper any other boosts they get.
I think a couple point bonus, or maybe some percentage bonus (Value x 1.25 or something like that) might fix the problem and put Ubers even more into the realm they belong.
Just something to think about. Might solve the problem.
I just didn't take HP into account under the assumption everyone has access to it. Same analysis as the items in terms of the spread.
Once you get to look at the equations you might see that it might not make too much of a difference in taking it into account (it would only boost the average power of the Special Attack Movepool value). Even so, I still have wifi in mind as well, where it would be hard to get a perfect 70 of the HP you want in order to really take it into account.
Anyway, I'm looking forward to more insight. The more help the better.
I'll let you decide if that's good or not.The more contribution, the better this system will be. So I'm gladd to hear your thoughts.
I guess part of the problem is that some of us are thinking in terms of Wifi. There's really no easy way to implement such a system over wifi, and you don't want people to have to tip their hand in advance. I think assigning values to items is the best way to go, it's just a question of implementation. Not to mention that a set number increase will more drastically change lower valued pokemon and barely dent upper valued ones (unless it is a percentage change of the base value). An item valued at 3 would boost a pokemon with a value of 6 by 50%, but would boost a pokemon with a value of 15 only 20% (if my math is right). Seems like it should be the other way around.
Another option, albeit a drastic one, is to go the Battle Tower route and limit a battle by not allowing more than one of any item. So, you'd have to choose where the leftovers go. Still chance for cheating, but less in my opinion. Easier to find two of the same item than trying to guess and do the math in your head.
I think items can be revisited later, but good thoughts all around.
I think I agree with TFP, even though it does seem counterproductive. I've lurked over at smogon, and I just get the general sense that they won't be very accpeting of the ideas presented here.
To be honest, I'm one guy that is using free time to do this for fun. I hit quite a bit of criticism right off the bat, and most of it was along the lines of "it can't be done and don't even try" rather than "I don't think this part of the math is accurate." I think most of the "constructive criticism" would fall under the first category if this was presented over at smogon. Particularly because everyone has their own opinion on which pokemon rocks, so to speak.
That aside, I do envision publicizing this more in the future for sure. I just think people go crazy when this thing is still in its infancy. They want it perfect right out of the gate or not at all, which doesn't make sense to me. So, I guess I want to make sure it is "presentable" before it goes further.
Although I wouldn't be opposed to someone bringing it up over there and directing them here
Now it's your turn to reach in my brain before I get there
I saw the Salamence discussion over at smogon. I agree with your analysis, and lots of people over there seem to as well in regards to the fact that there will always be a "most powerful" pokemon and it will be never ending if you keep kicking them out.
Salamence WAS 22 (216)...before the Stealth Rock modifier I added in to the mix. There was a little discussion about trying to make the modifier less drastic, which might help. Need to look at that more.
Still, I had almost the exact same thought about base stats where there should be a bonus for super high stats. I had an idea of maybe adding a "Total Base Stat" bonus. It would be a general way of giving a bonus for pokemon that would have a stat spread where several stats are high.
So, I just picked a number and went from there. I figured 100 or higher in multiple stats would pick out pokemon like you are describing, so spread across all six stats that would give a base stat total of 600 (averaging it out to account for stats above and below that line on a pokemon). I took a look at the list at Serebii and that looks like a perfect number. It would give a bonus to the following pokemon and no others:
Arceus, Mewtwo, Lugia, Ho-oh, Rayquaza, Dialga, Palkia, Giratina, Slaking, Kyogre, Groudon, Regigigas, Dragonite, Mew, Tyranitar, Celebi, Salamence, Metagross, Latias, Latios, Jirachi, Deoxys, Garchomp, Heatran, Cressalia, Manaphy, Darkrai, and Shaymin.
Not only would that further boost Ubers (which most of those are), but it will give a boost to some of the other "borderline" type pokemon like Salamence. There are a few odd ones like Slaking and Regigigas, but they shouldn't be really affected too much because their abilities will severely hamper any other boosts they get.
I think a couple point bonus, or maybe some percentage bonus (Value x 1.25 or something like that) might fix the problem and put Ubers even more into the realm they belong.
Just something to think about. Might solve the problem.
I just didn't take HP into account under the assumption everyone has access to it. Same analysis as the items in terms of the spread.
Once you get to look at the equations you might see that it might not make too much of a difference in taking it into account (it would only boost the average power of the Special Attack Movepool value). Even so, I still have wifi in mind as well, where it would be hard to get a perfect 70 of the HP you want in order to really take it into account.
Anyway, I'm looking forward to more insight. The more help the better.
RhythmScript
New Member
Not so much Smogon, but the two or three people AT Smogon who are responsible for Shoddy Battle 2.0 ... Smogon may have the largest Shoddy Servers but there are certainly others, and you don't have to get Smogon to change at all.
It's an addition, after all, not a replacement. Shoddy is set up to allow its battlers to do a LOT more than just what Smogon's standard rules allow; configuring a Shoddy server to have a max team PV wouldn't be hard to do at all... assuming you could get the PVs included in Shoddy's data table.
But even without PVS limit battling, having access to PVs is still a fairly useful thing, because it's essentially a numeric indicator of threat level. Especially since (like it or not) the litmus test for this system's efficacy is whether or not it tends to adhere to the general trends and classes of the Smogon tiers, having an on-screen visual indicator of the generally recognized "threat level" of a poke's PV while building it in Shoddy might be useful.
I don't see you getting "laughed out of town" because you're not really asking SMOGON to do anything different, just Shoddy (which is HQ'd on the Smogon forums). Worst-case scenario is they say "sorry, PVS is too much of a niche for us to cater to when developing SB2, we've already got our work cut out for us" in which case we accept it and move on.
FWIW I've been talking to Nomad (one of the Shoddy 1 devs; he developed the auto-team-builder IIRC) on PokemonGTS about my reservations with the Smogon tiering system and I have every intention of bringing this up with him, though I don't believe he's currently involved in the Shoddy coding process
As for Salamence and other similar pokes, I think the solution could be to weight any stat which is above average to account for that poke's ability to utilize that stat as a threat. Right now, the system accounts for regular base stats from the (probably fair) assumption that most pokes are built for one thing, such as physical sweeping. A pokemon only becomes a multi-threat if its stats are organized so that you could choose to dump EVs into one of two contradictory stats and either one would be a viable choice.
Bulbapedia, on this page, has a handy list of all the base stat averages, both for pokemon and for fully-evolved pokemon. I suppose it's up to you guys and some trial and error which should be used.
But for example, the average base attack stat for all fully-evolved pokes is 90, and the average base sp.atk stat is 85. I guess logic would dictate that any poke whose base stat in either of these fields is above average would get it weighted; though that would mean proportionally punishing ANYTHING which is even one point above normal, and that seems counterintuitive. I guess you could take it as a percentage; a pokemon whose base stat in any particular field is 10%, 15%, whatever% ABOVE average gets it weighted. This way, the more stats a pokemon has in the above-average category (and, in theory, the more stats that pokemon can exploit for a certain function) the more disproportionately large its PV will be.
So in the instance of Salamence, because it has a base Atk of 135 and a base SpAtk of 110, both of these values would get weighted for being too far above average, thus increasing its value. Then, a pokemon which has the same attack stat and base stat totals (like Metagross) will still wind up with a lower PV than Mence because its stat distribution prevents it from having more than one attacking function, thus making it more predictable (and therefore less dangerous in the metagame).
This would pan out well for checking pokes like Arceus, who has fairly well above-average stats in ALL categories, as well. It would also help reign in the (IMO) overweighting of a poke like Jirachi, whose 100 base stats in all fields prevent it from being extremely effective at ANYTHING, offensively or defensively, despite having a 600 base stat total.
The issues with this approach are twofold. One thing that worries me is the shelf that will occur from this system, where a poke with something like 125 base stats "makes the cut" where a poke with 130 doesn't, thus making two pokes with almost equal attacking power have vastly different threat levels because of the arbitrary line. In that case, it might make more sense to add a gradient to the weighting so that the farther above average a poke is, the greater the weighting (sort of like an income tax), preventing that shelf.
The other thing that I'm unsure about is which stats should get this treatment. For example, should a pokemon with excellent Speed and Def get equal weighting as one with excellent Def and SpDef, despite the fact that a choice between Speed and Def on a pokemon used for walling is hardly as up-in-the-air as a choice between Def and SpDef on a pokemon used for walling? And what will happen to physical tank pokes, which have above average attack defense, but may be lacking something such as speed? These pokes don't pose as great a threat as one which have above average Atk and Sp.Atk because you're not going to upset the match figuring out whether it's attack or defense they've boosted.
So in that regard, maybe a poke should only get extra weighting if it has above average Atk AND Sp.Attack, or above average Def AND Sp.Def, because it can be used to fill one of two crucial roles. (But then again, if a poke has solid defense and attack, like Metagross for example, one might lose a poke trying to determine if it's being played as a wall with maxed out defense and HP or as a tank with maxed out attack and defense. I guess that's still less threatening than Mence or Chomp, but more threatening than, say, Blissey, if only because a Gross may be designed to absorb attacks or to hit back. And then you're toeing the line of "how much can this point system reveal before it's impossible to account for human strategy?")
EDIT: As for Hidden Power, I may be overstating it's importance in type coverage as a whole. Don't doubt that there are people out there who will IV breed specifically for a certain Hidden Power type, though. Not just to add arbitrary grass type coverage to a poke, obviously, but because HP offers you the unique advantage of teaching your pokemon a type of move it would otherwise be incapable of learning.
The classic case is Magnezone; its current success story in the metagame is pretty much exclusively as a counter to Scyther. But being a Steel/Electric typing, it can't hit Scyther for Supereffective damage with much of anything; except that it's ability to learn a fire-type move (HP Fire) makes it suddenly far more valuable.
In fact, part of Magenzone's biggest use is the fact that magnet pull allows it to trap steel types, while HP Fire allows it to pretty much always hit for super-effective damage.
Now, it's true that any pokemon can learn hidden power, but what benefits does a pokemon like Gyarados see from it? you'll never see Gyara with an HP equipped because, for what it's meant to do, it doesn't need that type coverage.
I guess my point is, in terms of real threat level, a pokemon's type coverage only matters (or matters much MORE) insofar as it's ability to hit for supereffective damage against the types which can hit IT for supereffective damage; logic dictates a psychic type pokemon with a kickass movepool of every type except dark and fighting is far LESS valuable than a comparable psychic type with a kickass fighting and dark type in its movepool because Dark and Ghost are the types it's going to be most at risk against. Hidden Power is irrelevant on pokes which already have good movepool type coverage to deal with the pokes that can take them out, but it provides an invaluable (and, by your math, ignored) boost in threat level on a poke which may have trouble with one or two types (like a steel type using Ice HP to help deal with the ground type pokes hitting it with STAB Earthquake)
It's an addition, after all, not a replacement. Shoddy is set up to allow its battlers to do a LOT more than just what Smogon's standard rules allow; configuring a Shoddy server to have a max team PV wouldn't be hard to do at all... assuming you could get the PVs included in Shoddy's data table.
But even without PVS limit battling, having access to PVs is still a fairly useful thing, because it's essentially a numeric indicator of threat level. Especially since (like it or not) the litmus test for this system's efficacy is whether or not it tends to adhere to the general trends and classes of the Smogon tiers, having an on-screen visual indicator of the generally recognized "threat level" of a poke's PV while building it in Shoddy might be useful.
I don't see you getting "laughed out of town" because you're not really asking SMOGON to do anything different, just Shoddy (which is HQ'd on the Smogon forums). Worst-case scenario is they say "sorry, PVS is too much of a niche for us to cater to when developing SB2, we've already got our work cut out for us" in which case we accept it and move on.
FWIW I've been talking to Nomad (one of the Shoddy 1 devs; he developed the auto-team-builder IIRC) on PokemonGTS about my reservations with the Smogon tiering system and I have every intention of bringing this up with him, though I don't believe he's currently involved in the Shoddy coding process
As for Salamence and other similar pokes, I think the solution could be to weight any stat which is above average to account for that poke's ability to utilize that stat as a threat. Right now, the system accounts for regular base stats from the (probably fair) assumption that most pokes are built for one thing, such as physical sweeping. A pokemon only becomes a multi-threat if its stats are organized so that you could choose to dump EVs into one of two contradictory stats and either one would be a viable choice.
Bulbapedia, on this page, has a handy list of all the base stat averages, both for pokemon and for fully-evolved pokemon. I suppose it's up to you guys and some trial and error which should be used.
But for example, the average base attack stat for all fully-evolved pokes is 90, and the average base sp.atk stat is 85. I guess logic would dictate that any poke whose base stat in either of these fields is above average would get it weighted; though that would mean proportionally punishing ANYTHING which is even one point above normal, and that seems counterintuitive. I guess you could take it as a percentage; a pokemon whose base stat in any particular field is 10%, 15%, whatever% ABOVE average gets it weighted. This way, the more stats a pokemon has in the above-average category (and, in theory, the more stats that pokemon can exploit for a certain function) the more disproportionately large its PV will be.
So in the instance of Salamence, because it has a base Atk of 135 and a base SpAtk of 110, both of these values would get weighted for being too far above average, thus increasing its value. Then, a pokemon which has the same attack stat and base stat totals (like Metagross) will still wind up with a lower PV than Mence because its stat distribution prevents it from having more than one attacking function, thus making it more predictable (and therefore less dangerous in the metagame).
This would pan out well for checking pokes like Arceus, who has fairly well above-average stats in ALL categories, as well. It would also help reign in the (IMO) overweighting of a poke like Jirachi, whose 100 base stats in all fields prevent it from being extremely effective at ANYTHING, offensively or defensively, despite having a 600 base stat total.
The issues with this approach are twofold. One thing that worries me is the shelf that will occur from this system, where a poke with something like 125 base stats "makes the cut" where a poke with 130 doesn't, thus making two pokes with almost equal attacking power have vastly different threat levels because of the arbitrary line. In that case, it might make more sense to add a gradient to the weighting so that the farther above average a poke is, the greater the weighting (sort of like an income tax), preventing that shelf.
The other thing that I'm unsure about is which stats should get this treatment. For example, should a pokemon with excellent Speed and Def get equal weighting as one with excellent Def and SpDef, despite the fact that a choice between Speed and Def on a pokemon used for walling is hardly as up-in-the-air as a choice between Def and SpDef on a pokemon used for walling? And what will happen to physical tank pokes, which have above average attack defense, but may be lacking something such as speed? These pokes don't pose as great a threat as one which have above average Atk and Sp.Atk because you're not going to upset the match figuring out whether it's attack or defense they've boosted.
So in that regard, maybe a poke should only get extra weighting if it has above average Atk AND Sp.Attack, or above average Def AND Sp.Def, because it can be used to fill one of two crucial roles. (But then again, if a poke has solid defense and attack, like Metagross for example, one might lose a poke trying to determine if it's being played as a wall with maxed out defense and HP or as a tank with maxed out attack and defense. I guess that's still less threatening than Mence or Chomp, but more threatening than, say, Blissey, if only because a Gross may be designed to absorb attacks or to hit back. And then you're toeing the line of "how much can this point system reveal before it's impossible to account for human strategy?")
EDIT: As for Hidden Power, I may be overstating it's importance in type coverage as a whole. Don't doubt that there are people out there who will IV breed specifically for a certain Hidden Power type, though. Not just to add arbitrary grass type coverage to a poke, obviously, but because HP offers you the unique advantage of teaching your pokemon a type of move it would otherwise be incapable of learning.
The classic case is Magnezone; its current success story in the metagame is pretty much exclusively as a counter to Scyther. But being a Steel/Electric typing, it can't hit Scyther for Supereffective damage with much of anything; except that it's ability to learn a fire-type move (HP Fire) makes it suddenly far more valuable.
In fact, part of Magenzone's biggest use is the fact that magnet pull allows it to trap steel types, while HP Fire allows it to pretty much always hit for super-effective damage.
Now, it's true that any pokemon can learn hidden power, but what benefits does a pokemon like Gyarados see from it? you'll never see Gyara with an HP equipped because, for what it's meant to do, it doesn't need that type coverage.
I guess my point is, in terms of real threat level, a pokemon's type coverage only matters (or matters much MORE) insofar as it's ability to hit for supereffective damage against the types which can hit IT for supereffective damage; logic dictates a psychic type pokemon with a kickass movepool of every type except dark and fighting is far LESS valuable than a comparable psychic type with a kickass fighting and dark type in its movepool because Dark and Ghost are the types it's going to be most at risk against. Hidden Power is irrelevant on pokes which already have good movepool type coverage to deal with the pokes that can take them out, but it provides an invaluable (and, by your math, ignored) boost in threat level on a poke which may have trouble with one or two types (like a steel type using Ice HP to help deal with the ground type pokes hitting it with STAB Earthquake)
Like I said before, a 25% reduction from a SR weak is too much.
In competitive battling, SR double weak (50%) is severely crippling while a SR 25% weak isn't that much.
Either use my single weak = 6/7 (-14.3%) and double weak = 4/7 (-42.9%) or your single weak = 7/8 (-12.5%) and double weak = 5/8 (-37.5%). Both sets of numbers reflect the fact that double weak is crippling while single weak isn't.
By the way, my SR calculation is multiplicative, multiplying the reduction by 8/7 to factor out the average 7/8 reduction, while yours is additive, adding 1/8 to the reduction to factor it out. Use whichever seems to work better.
With my values, Salamence becomes 19 (185), and with your values, Salamence becomes 19 (189). The difference is quite small, but might make a difference.
EDIT: Okay, I just noticed that I quintuple posted this by accident...darn database errors...
EDIT 2:
Yes. Just add a few spaces to the end of each Pokemon's name that has their numbers shifted too far left. After adding enough spaces, the numbers will align with the rest.
RhythmScript
New Member
The other issue (beyond the Stealth Rock modifier overdoing it by quite a bit) is the fact that there seems to be some disparity in the scaling of the numbers. Most of the focus, I think, has been on getting the pokes to appear in the "right order" in the threat list, so that their percieved level of danger tends to correspond to their place. But whether or not the pokes themselves are ranked correctly, I think the gaps between them might be too large.
For shits and giggles I decided to look up the values of my Steel Gym team. All OU, and I didn't even put Heatran on the list (though Jirachi, who IMO is still grossly overvalued by this system, is present). Here are the numbers:
Metagross - 187
Scizor - 160
Magnezone - 151
Bronzong - 155
Skarmory - 183
Jirachi - 265
TOTAL -- 1101
A 1,101 value team? That's insane. If you do the rounding instead, you get a 111 team; so it's slightly lower when you count in the hundreds, but not much. I see people pulling for the 70-75 range to keep it OU/UU mixed, and I think that's great, but that should mean that 80-90 would yield a standard OU set... and here, an all OU set is pulling down numbers that are 20-30 points HIGHER than expected.
In no small part, I think this is due to Jirachi's overly high rating, which could be balanced out by adjusting the formula to weight stats by some percentile once they were above 100... say, for every point beyond 100 in that stat, a poke gets an additional 1% of its total value added to its stat? So a poke at 100 or less stats gets taken at 100%... a poke at 130 gets 130% OF 130 (169) a poke at 115 gets 115% OF 115 (132), etc. This would help to make "useful stats" count for more on the overall scale, thus decreasing (by relation) the value of pokes like Jirachi with no truly "applicable" stats but a high base stat average, while simultaneously punishing pokes like Salamence and Garchomp whose multiple exceptional stats make them unpredictable and dangerous.
I should note that this math is a) scalar (if it makes for too dramatic of an adjustment, simply step 1% per point down to 0.5% per point. The cutoff of 100 can be moved up or down as well to avoid catching well-placed pokes in its net), and b) may be applied only to certain stats (It almost certainly doesn't matter for the speed stat, and it's probably more important on the attacking stats than the defensive stats. But that depends, I suppose, on which pokemon seem most drastically affected by this approach).
Back on topic, I think another reason the team is valued so highly (and this may or may not be fixed by the above stats) is I think resistances are being counted too strongly. This is most obvious with steel type pokemon, because despite their almost total lack of supereffective STAB, they have incredible resistances and immunities to almost every type across the board (As a rule steel is the only type with more immunities/resistances than neutral/supereffectives, I think). It's especially apparent with a Poke like Skarmory, whose value in a team isn't really very much except as a utility poke (there are far better steel type walls and he's really only helpful as a spiker/whirlwinder). I'm not saying Skarmory isn't good, but for him to be valued above Scizor is insane. Even if the OU usage trends, with Scizor appearing like 1 in 6 matches, are ignored.... Scizor is capable of holding his own in Ubers with STAB tech bullet punch and his high base attack, Skarmory would stand literally NO chance in Ubers.
The root of this issue, I think, is how good sweeping ability completely preempts a need for good defenses. A pokemon with sufficiently high attack and speed (or, in the case of Scizor, sufficiently high priority attack) can get by with essentially NO defenses because it can eliminate most threats without fearing being hit. Meanwhile, Skarmory has higher defense than Scizor has offense, but because its defense/HP isn't high ENOUGH to make it truly useful as a wall compared to pokes like Blissey and Snorlax, it has to take on a role as a utility poke. But even WITH Scizor's tech/bulletpunch modifier, it winds up ranked lower, presumably because Skar's typing and defenses make it seem like a much more successful wall than it is.
What concerns me is the fact that Skarmory's defenses are higher than Scizor's offenses, meaning if a weighted system were introduced, the gap between them would actually GROW, rather than shrinking, despite the above stated reality that Scizor is just a far more powerful poke in general than Skar.
Perhaps the issue is HP's usage in the formula; it seems to be treated as a modifier for defense, much as speed is a modifier for offense. But the most successful wall/stall/defensive pokemon actually have higher base HP stats than defensive stats, so maybe HP needs to get more weight in the defensive side of the formula and the actual Def/SpDef stats need to get less. It may also be worth weighting the entire attack side of the formula to matter more than defense, if for no reason other than the movepool modifiers don't seem to give quite the overall bump to the PV that the type-resistance modifiers do, and the formula may be inadvertently increasing the weight of pokemon with good defensive roles and decreasing the weight of those with good offensive roles. When Blissey has a higher PV than Garchomp and Tyranitar, Jirachi has a higher PV than Heatran (and Garchomp and Tyranitar), and Chansey has a higher PV than Lucario, Salamence, Celebi, and Metagross, I think it's probably a reality that defense is being valued too greatly over offense.
Some food for thought
Kindrindra
大事なのは自分らしいくある事
Actually, if memory serves correct, that's one of the first things that was discussed about how to factor it in. So, ya, it is.
I like the idea for pokemon gaining a small percentage for every point over 100, rather than getting a set bonus for being at 100 or over.
Dragoon952
The Winter Moth
@Zowayix: I think your stealth rock modifier idea is better than the 25/50 I was doing. I'm going to look at it in more detail soon, but it makes much more sense. It is more correct because 1/8 HP damage is the baseline and not 0 HP damage.
I actually really didn't take into account getting things into the "right place" so to speak. I tried to find a way to adjust numbers that the game gives us and then see where pokemon land. Even though I didn't take usage and smogon into account for the most part, I did use it as a benchmark to know whether I'm on the right path. The numbers just turned out the way they did.
For this system to work, there has to be gaps, but I guess the question is how big the gaps should be. There has to be enough that "good" pokemon are relatively expensive and "bad" pokemon aren't in order to create an environment where there are consequences for shooting for the big guns.
Well, that's not entirely insane. The 70 point team value total which we've been discussing and testing give you between 11 and 12 points per pokemon on average, and the 10-12 range seems to be the middle ground for all pokemon (kind of the border between OU and UU, so to speak). We're just finding that is giving us a good mixed environment because choosing a "better" OU pokemon is forcing the choice of a more UU pokemon to get back to the average 10-12 range between the two. We're just finding that is producing a good OU/UU mixed battle.
That aside, I have NO idea what a good OU number would be. An average of 15 per would give you a total value of 90, but OU seems to mostly top out at around 20 points per pokemon (120) total. The closer you get to that top range, you are basically just reverting to standard OU smogon rules, so our focus has been mainly on testing the border numbers to see what happens.
Still, the thing I wanted to get at was getting away from the smogon notion that an OU pokemon is an OU pokemon, so they are all basically equal. There is definitely a range, with some OU better than others. I wanted to bring that out somehow. And that brings us to your team, which, if you think about it (Jirachi aside), isn't Skarm one of the "best" defensive pokemon, along with Bronzong? Isn't Scizor one of the better offensive pokemon in the game? Isn't Metagross up there as well? They are all Steel, so it could be a problem with how the modifiers are calculated. But, if the system is working, it could be because your team has a lot of top level pokemon.
And that's what we're here to figure out I'll go into my methodology later so we can try and figure this problem out.
If you look at Jirachi's numbers, he gets a great deal of value from his support movepool and his ability (in terms of how I use it in calculation). Jirachi is highly rated then because of his versatility and not his focus on one particular stat. Garchomp gets most of his value from offense and Blissey from defense. Jirachi gets value from being unpredicatable, basically. And that goes to my methodology which we can try to analyze and see where things are getting mucked up.
I like your idea, but it just appears that something else would need to be applied as well if we use it. Boosting certain pokemon with certain high stats further will broaden the gaps with the lower level guys, and it may need to have some other adjustment to keep the gaps reasonable. We can figure that out too.
I guess it would help to explain my methodology a bit so we can get at some issues.
I basically figured that, based upon analyzing how pokemon are used, there are three different functions: attackers, defenders, and support. So I knew I wanted to divide the categories up in that kind of fashion.
I started off by trying to get all of the known numbers that pokemon give us by looking at how damage is calculated (which is the heart of pokemon for the most part). When you take out all of the random variables, you are basically left with attack stat + attack strength vs. defense. So I started there. And that covers four of the six stats in terms of application (whatever application that may be).
HP and Speed seemed different to me, because their value is derived from how they affect other stats rather than in of themselves. HP defines how long you can survive taking hits with your defenses, and speed seemed important to a general ability to strike first. So, I wanted to use them as a "modifier" rather than a direct number.
The whole idea in this system is not "how will this pokemon do against a threat in its tier," but rather "how would this pokemon do, out of the gate, against a random opponent." It is basically comparing the pokemon to the entire range of opponents, with an above average rating showing that it will fair better more often than not against a random opponent and below average means it will struggle more often than not.
I used that as a basis for the HP/Speed modifiers, by taking the stat against the median of all pokemon in that category (the quintessential marker for equal number above and below). It turned out to be 65 for both, so I used HP and Speed by comparing it to the median in terms of how much it deviates from it. Then, that can be used to modify defenses and offenses. In essence, it is saying that a pokemon who has the same attack but is faster than another pokemon will do better.
Attack and defense I used just like the damage formula. The main difference is STAB, so I wanted to separate that from the rest (because non-stab is the same for everyone in terms of relative power if the stats are the same). So that's why I took the "most powerful STAB" approach to account for the most usable move and not worry about all of the weaker STAB moves that would never be used competitively.
The last basic function in terms of offense and defense I wanted to use as a modifier was type. In theory, a base 100 Dragon attack is more effective that a base 100 Poison attack when facing a random opponent.
The way I went about that was comparing good vs. bad in terms of typing. I used 1 for neutral, since a modifier of 1 wouldn't change any value as a multiplicative function. I then used 2 for 2x effectiveness and 4 as 4x effectiveness. It then became the basic equation of (Good + Neutral) / (Bad + Neutral) to give a modifier of it's effectiveness. Above 1 would mean that more often than average you will have a super effective move available, whereas defensively it would mean more often than average you'd have a resistance to an attack. Immunity had to be taken into account, but it had to be more subjective. Since it is better/worse than a 4x effectiveness, I gave it a value of 5 too not try and overweight it.
This gave a range offensively of 0.34 (Poison) to 1.24 (Rock). However, offensively is a completely different story. Since you can only attack with one type, it was a much bigger pool of values because you are taking ALL typing (including dual typing) into account when determining offensive effectiveness. Conversely, defensively only needed to take into account each attacking type (which is a pool of 17 values), and this gave a range of 0.62 (Ice/Psychic) up to 4.57 (Dragon/Steel). Most Steel types, because of their massive resistances, give them defensive modifiers of around 3 and up, which is a very large difference in comparison than what offensively they get.
So, taking that into account, you could argue that this isn't taking typing into account effectively enough. Which is completely possible. Perhaps averaging it out would be better than comparing (assuming my values are correct in taking things into account). Still, it is possible that Defenses will be generally rated higher in terms of typing than offenses as a modifier. There is an underlying reason for this in the fact there is no offensive equivalent of an "immunity." It would be like saying everything is a one hit KO if you go to the opposite end of the spectrum. So, because of that, defense will always have an advantage in terms of typing as a modifier.
Anyway, this is all long winded. Might be easier to discuss all of that before even considering movepool. Feel free to comment/suggest if you see some better ways of doing things.
I actually really didn't take into account getting things into the "right place" so to speak. I tried to find a way to adjust numbers that the game gives us and then see where pokemon land. Even though I didn't take usage and smogon into account for the most part, I did use it as a benchmark to know whether I'm on the right path. The numbers just turned out the way they did.
For this system to work, there has to be gaps, but I guess the question is how big the gaps should be. There has to be enough that "good" pokemon are relatively expensive and "bad" pokemon aren't in order to create an environment where there are consequences for shooting for the big guns.
Well, that's not entirely insane. The 70 point team value total which we've been discussing and testing give you between 11 and 12 points per pokemon on average, and the 10-12 range seems to be the middle ground for all pokemon (kind of the border between OU and UU, so to speak). We're just finding that is giving us a good mixed environment because choosing a "better" OU pokemon is forcing the choice of a more UU pokemon to get back to the average 10-12 range between the two. We're just finding that is producing a good OU/UU mixed battle.
That aside, I have NO idea what a good OU number would be. An average of 15 per would give you a total value of 90, but OU seems to mostly top out at around 20 points per pokemon (120) total. The closer you get to that top range, you are basically just reverting to standard OU smogon rules, so our focus has been mainly on testing the border numbers to see what happens.
Still, the thing I wanted to get at was getting away from the smogon notion that an OU pokemon is an OU pokemon, so they are all basically equal. There is definitely a range, with some OU better than others. I wanted to bring that out somehow. And that brings us to your team, which, if you think about it (Jirachi aside), isn't Skarm one of the "best" defensive pokemon, along with Bronzong? Isn't Scizor one of the better offensive pokemon in the game? Isn't Metagross up there as well? They are all Steel, so it could be a problem with how the modifiers are calculated. But, if the system is working, it could be because your team has a lot of top level pokemon.
And that's what we're here to figure out
I'll go into my methodology later so we can try and figure this problem out.If you look at Jirachi's numbers, he gets a great deal of value from his support movepool and his ability (in terms of how I use it in calculation). Jirachi is highly rated then because of his versatility and not his focus on one particular stat. Garchomp gets most of his value from offense and Blissey from defense. Jirachi gets value from being unpredicatable, basically. And that goes to my methodology which we can try to analyze and see where things are getting mucked up.
I like your idea, but it just appears that something else would need to be applied as well if we use it. Boosting certain pokemon with certain high stats further will broaden the gaps with the lower level guys, and it may need to have some other adjustment to keep the gaps reasonable. We can figure that out too.
I guess it would help to explain my methodology a bit so we can get at some issues.
I basically figured that, based upon analyzing how pokemon are used, there are three different functions: attackers, defenders, and support. So I knew I wanted to divide the categories up in that kind of fashion.
I started off by trying to get all of the known numbers that pokemon give us by looking at how damage is calculated (which is the heart of pokemon for the most part). When you take out all of the random variables, you are basically left with attack stat + attack strength vs. defense. So I started there. And that covers four of the six stats in terms of application (whatever application that may be).
HP and Speed seemed different to me, because their value is derived from how they affect other stats rather than in of themselves. HP defines how long you can survive taking hits with your defenses, and speed seemed important to a general ability to strike first. So, I wanted to use them as a "modifier" rather than a direct number.
The whole idea in this system is not "how will this pokemon do against a threat in its tier," but rather "how would this pokemon do, out of the gate, against a random opponent." It is basically comparing the pokemon to the entire range of opponents, with an above average rating showing that it will fair better more often than not against a random opponent and below average means it will struggle more often than not.
I used that as a basis for the HP/Speed modifiers, by taking the stat against the median of all pokemon in that category (the quintessential marker for equal number above and below). It turned out to be 65 for both, so I used HP and Speed by comparing it to the median in terms of how much it deviates from it. Then, that can be used to modify defenses and offenses. In essence, it is saying that a pokemon who has the same attack but is faster than another pokemon will do better.
Attack and defense I used just like the damage formula. The main difference is STAB, so I wanted to separate that from the rest (because non-stab is the same for everyone in terms of relative power if the stats are the same). So that's why I took the "most powerful STAB" approach to account for the most usable move and not worry about all of the weaker STAB moves that would never be used competitively.
The last basic function in terms of offense and defense I wanted to use as a modifier was type. In theory, a base 100 Dragon attack is more effective that a base 100 Poison attack when facing a random opponent.
The way I went about that was comparing good vs. bad in terms of typing. I used 1 for neutral, since a modifier of 1 wouldn't change any value as a multiplicative function. I then used 2 for 2x effectiveness and 4 as 4x effectiveness. It then became the basic equation of (Good + Neutral) / (Bad + Neutral) to give a modifier of it's effectiveness. Above 1 would mean that more often than average you will have a super effective move available, whereas defensively it would mean more often than average you'd have a resistance to an attack. Immunity had to be taken into account, but it had to be more subjective. Since it is better/worse than a 4x effectiveness, I gave it a value of 5 too not try and overweight it.
This gave a range offensively of 0.34 (Poison) to 1.24 (Rock). However, offensively is a completely different story. Since you can only attack with one type, it was a much bigger pool of values because you are taking ALL typing (including dual typing) into account when determining offensive effectiveness. Conversely, defensively only needed to take into account each attacking type (which is a pool of 17 values), and this gave a range of 0.62 (Ice/Psychic) up to 4.57 (Dragon/Steel). Most Steel types, because of their massive resistances, give them defensive modifiers of around 3 and up, which is a very large difference in comparison than what offensively they get.
So, taking that into account, you could argue that this isn't taking typing into account effectively enough. Which is completely possible. Perhaps averaging it out would be better than comparing (assuming my values are correct in taking things into account). Still, it is possible that Defenses will be generally rated higher in terms of typing than offenses as a modifier. There is an underlying reason for this in the fact there is no offensive equivalent of an "immunity." It would be like saying everything is a one hit KO if you go to the opposite end of the spectrum. So, because of that, defense will always have an advantage in terms of typing as a modifier.
Anyway, this is all long winded. Might be easier to discuss all of that before even considering movepool. Feel free to comment/suggest if you see some better ways of doing things.
RhythmScript
New Member
My point was, though, that you shouldn't use HP as a modifier for defensive stats, because in reality it is THE primary defensive stat. A poke which is being EV trained for bulk, for instance, will always get the EVs to HP first, and then to defenses second (Bronzong, for example is recommended 252 in HP, and then a spread in Sp. Def and Def or Atk).
I think your math treats defensive stats as more important than HP, when in reality I'd say the opposite is true. HP is the most important of the three "survival stats"
Beyond that your math looks good, but I guess what I'm saying is you should just weight the attack side of the formula to compensate for the fact that there are so many modifiers on defense (like immunities). I think the immunities math tends to bump up the defensive side of the equation, while there is no similar bump to the attack side. Which, again, is why defensively-oriented pokemon seem to be ranked higher than offensively-oriented pokemon. Which makes sense at first glance, but really punishes offensively-oriented pokes with generally irrelevantly good defenses (like Scizor, whose main kickass-factor is that he has such a powerful priority move).
I think your math treats defensive stats as more important than HP, when in reality I'd say the opposite is true. HP is the most important of the three "survival stats"
Beyond that your math looks good, but I guess what I'm saying is you should just weight the attack side of the formula to compensate for the fact that there are so many modifiers on defense (like immunities). I think the immunities math tends to bump up the defensive side of the equation, while there is no similar bump to the attack side. Which, again, is why defensively-oriented pokemon seem to be ranked higher than offensively-oriented pokemon. Which makes sense at first glance, but really punishes offensively-oriented pokes with generally irrelevantly good defenses (like Scizor, whose main kickass-factor is that he has such a powerful priority move).
Dragoon952
The Winter Moth
Offense naturally gets a bump because it has two things it takes into account: Base Stat and Attack Strength. Defenses only take into account Base Stat so naturally are lower from the beginning. The difference seems to be the type modifier. Steel types are just in a class of their own and there is no true offensive equivalent to what Steel types can do defensively. Not sure how to counter balance that. Open to suggestions if you think Steels are overrated. I usually only see drastically higher defensive ratings for highly defensive types (Skarmory, Blissey, etc.) or if there is poor offensive STAB type coverage (like Grass and Poison monotypes).
As for the EV, I'm not entirely sure of this, but doesn't a pokemon take less damage percentage wise when maxing out a defensive stat rather than HP? I think the EV spreads are the way they are more often than not because you don't know if you'll be dealing with a special or physical attack. But, I thought straight up, raising Defense or Special Defense will allow longer surivial rather than raising HP. I could be wrong though.
Steelix211
New Member
You're right about the HP EV spreads Dragoon; HP is added to most often to help to take physical and special hits better. I know for a fact that many Pokemon benefit more from maximising the respective Defense over HP, I was running damage calcs on a Mew a while back. I think it depends more on the Pokemon's ratio of HP to defenses though. Pokemon like Spiritomb would benefit from HP whilst Wailord would benefit more from Defenses.
Steelix211
New Member
*BUMP*
Also, just had the most intense tryout battle with TheFightingPikachu. 70 seems like a really good number to have matches with; I was constantly on my toes wondering if the next opponent would be OU, UU or even NU. Kudos to everyone who's contributed to the system so far - I'm definitely looking forward to more matches like this!
Also, just had the most intense tryout battle with TheFightingPikachu. 70 seems like a really good number to have matches with; I was constantly on my toes wondering if the next opponent would be OU, UU or even NU. Kudos to everyone who's contributed to the system so far - I'm definitely looking forward to more matches like this!
Dragoon952
The Winter Moth
Good stuff. Glad it's working.
I've been busy lately, but I'll pick it up again soon. I need to test the different stealth rock modifier numbers that were presented, see if you guys like the results, and then continue plowing through NU.