-
Welcome to Smogon! Take a moment to read the Introduction to Smogon for a run-down on everything Smogon, and make sure you take some time to read the global rules.
-
Congrats to the winners of the 2023 Smog Awards!
Doubles Ladder + Seasonal Stats
- Thread starter Level 51
- Start date
I feel like this data implies that people are bringing a good number of checks, even if not consciously, for the pokemon with bad w/l ratios.
it would be interesting to see a graph showing w/l ratio from one tour to the next; then we would see pokemon that dropped off b/c people brought checks, and pokemon that moved up because people didnt bring checks for them, and pokemon that say in the same position b/c they are so commonly used and well balanced like lando. the game seems to be pretty trendy
it would be interesting to see a graph showing w/l ratio from one tour to the next; then we would see pokemon that dropped off b/c people brought checks, and pokemon that moved up because people didnt bring checks for them, and pokemon that say in the same position b/c they are so commonly used and well balanced like lando. the game seems to be pretty trendy
cress is shit
Stratos
Banned deucer.
this is true, it's dropping on the VR as soon as Laga and Arcticblast vote or kamikaze17 replaces them (hint hint)
not below 2 tho cuz offensive cress is still a threat. cm is aids though
not below 2 tho cuz offensive cress is still a threat. cm is aids though
checkroom is 5-2 in this ssnl nb
maybe we could get some stats on mons' matchups against things not itself ¯\_(ツ)_/¯
Maybe you could try to do the fucking analysis instead of sitting around complaining that the stats are not good enough. I almost got a fucking migraine doing the first 1.5 seasonals' worth of stats, and Stratos hasn't had much fun compiling it either. It's harder than you think it is, especially when we're compiling over a thousand data points by hand.
Also so that this post isn't completely negative, thanks for doing the Fall Seasonals stats Stratos! I'll clean it up and add it to the OP in the days to come. n_n!
my post wasn't to be a complaining fuck, if that's the impression it gave I'm sorry. All I wanted was some dialogue on whether or not it was a possibility also shay said he'd share some of the stuff he did so n_n
if your offering the raw data, I would like to mess around with it
The raw data is freely available! Just go to the Seasonal thread for the corresponding round and w r i t e a l l t h e d a t a d o w n y o u r s e l f
wait so you look at replys and write down each persons team?
yes thats what is done. We dont have access to much beyond that
After stalling for almost a month, I'm proud to announce that I've taken Stratos' amazing work on the Fall Seasonal Stats and revamped it! The new stats spreadsheet (click!) comes with an all-new visualiser, and I feel fairly safe in saying that it presents tournament stats in a way that hasn't been done on Smogon before (or perhaps on much of the web). I hope you enjoy it, and feel free to drop me a comment here if you have any feedback regarding any part of the spreadsheet.
I've also added the spreadsheet to the OP. A belated Merry Christmas to everyone!
I've also added the spreadsheet to the OP. A belated Merry Christmas to everyone!
Forgive the double post, but I wanted to share this and I figured it'd get more attention if I posted it separately. Here (click!) are the teams, top 32 and up, from the Fall Seasonal—well, the ones that I could find, anyway. If you're more of a visual person, feel free to pop over here, where I've posted the same data but in image form.
Hi guys! I thought this would be a cool post, Level 51 gave me permission to do the teams used in SPL :x, also here are some usage stats that you might have missed (courtesy of McMeghan):
Here are all teams used, the L in front of a team implies the team has lost. The W in front of a team implies the team has won:
http://puu.sh/mKnH0/56f5f60f68.txt
However, if you like visual stuff more like me, you can click on these hide tags below!
I tried to follow Level 51's layout, but I couldn't get rid of the /'s between Pokemon without fucking up the formatting or having to do everything over again. Level 51 edit: I did this!
I hope you enjoy! n__n:
I will keep this thread updated with more SPL games and teams used! :D
Code:
+ ---- + ------------------ + --- + ------- + ------ +
| Rank | Pokemon | Use | Usage % | Win % |
+ ---- + ------------------ + --- + ------- + ------ +
| 1 | Landorus-Therian | 13 | 65.00% | 53.85% |
| 2 | Thundurus | 9 | 45.00% | 33.33% |
| 3 | Kangaskhan | 8 | 40.00% | 62.50% |
| 4 | Aegislash | 7 | 35.00% | 57.14% |
| 5 | Charizard | 6 | 30.00% | 33.33% |
| 6 | Amoonguss | 5 | 25.00% | 60.00% |
| 6 | Heatran | 5 | 25.00% | 60.00% |
| 8 | Kyurem-Black | 4 | 20.00% | 50.00% |
| 9 | Talonflame | 3 | 15.00% | 100.0% |
| 9 | Gengar | 3 | 15.00% | 66.67% |
| 9 | Jirachi | 3 | 15.00% | 66.67% |
| 9 | Landorus | 3 | 15.00% | 66.67% |
| 9 | Rotom-Wash | 3 | 15.00% | 66.67% |
| 9 | Venusaur | 3 | 15.00% | 0.00% |
| 15 | Azumarill | 2 | 10.00% | 100.0% |
| 15 | Weavile | 2 | 10.00% | 100.0% |
| 15 | Keldeo | 2 | 10.00% | 50.00% |
| 15 | Metagross | 2 | 10.00% | 50.00% |
| 15 | Sylveon | 2 | 10.00% | 50.00% |
| 15 | Terrakion | 2 | 10.00% | 50.00% |
| 15 | Zapdos | 2 | 10.00% | 50.00% |
| 15 | Gardevoir | 2 | 10.00% | 0.00% |
| 15 | Gyarados | 2 | 10.00% | 0.00% |
| 15 | Latios | 2 | 10.00% | 0.00% |
| 25 | Aerodactyl | 1 | 5.00% | 100.0% |
| 25 | Conkeldurr | 1 | 5.00% | 100.0% |
| 25 | Deoxys-Attack | 1 | 5.00% | 100.0% |
| 25 | Diancie | 1 | 5.00% | 100.0% |
| 25 | Ferrothorn | 1 | 5.00% | 100.0% |
| 25 | Infernape | 1 | 5.00% | 100.0% |
| 25 | Mew | 1 | 5.00% | 100.0% |
| 25 | Rhyperior | 1 | 5.00% | 100.0% |
| 25 | Salamence | 1 | 5.00% | 100.0% |
| 25 | Suicune | 1 | 5.00% | 100.0% |
| 25 | Tangrowth | 1 | 5.00% | 100.0% |
| 25 | Blastoise | 1 | 5.00% | 0.00% |
| 25 | Blaziken | 1 | 5.00% | 0.00% |
| 25 | Breloom | 1 | 5.00% | 0.00% |
| 25 | Chesnaught | 1 | 5.00% | 0.00% |
| 25 | Clefable | 1 | 5.00% | 0.00% |
| 25 | Cresselia | 1 | 5.00% | 0.00% |
| 25 | Entei | 1 | 5.00% | 0.00% |
| 25 | Excadrill | 1 | 5.00% | 0.00% |
| 25 | Hoopa-Unbound | 1 | 5.00% | 0.00% |
| 25 | Hydreigon | 1 | 5.00% | 0.00% |
| 25 | Porygon2 | 1 | 5.00% | 0.00% |
| 25 | Rhydon | 1 | 5.00% | 0.00% |
| 25 | Victini | 1 | 5.00% | 0.00% |
| 25 | Virizion | 1 | 5.00% | 0.00% |
+ ---- + ------------------ + --- + ------- + ------ +
Code:
+ ---- + ------------------ + --- + ------- + ------ +
| Rank | Pokemon | Use | Usage % | Win % |
+ ---- + ------------------ + --- + ------- + ------ +
| 1 | Landorus-Therian | 6 | 60.00% | 66.67% |
| 2 | Aegislash | 4 | 40.00% | 75.00% |
| 2 | Charizard | 4 | 40.00% | 25.00% |
| 4 | Kangaskhan | 3 | 30.00% | 66.67% |
| 4 | Thundurus | 3 | 30.00% | 33.33% |
| 6 | Amoonguss | 2 | 20.00% | 100.0% |
| 6 | Heatran | 2 | 20.00% | 100.0% |
| 6 | Kyurem-Black | 2 | 20.00% | 100.0% |
| 6 | Gengar | 2 | 20.00% | 50.00% |
| 6 | Jirachi | 2 | 20.00% | 50.00% |
| 6 | Landorus | 2 | 20.00% | 50.00% |
| 6 | Rotom-Wash | 2 | 20.00% | 50.00% |
| 6 | Latios | 2 | 20.00% | 0.00% |
| 6 | Venusaur | 2 | 20.00% | 0.00% |
| 15 | Aerodactyl | 1 | 10.00% | 100.0% |
| 15 | Azumarill | 1 | 10.00% | 100.0% |
| 15 | Conkeldurr | 1 | 10.00% | 100.0% |
| 15 | Diancie | 1 | 10.00% | 100.0% |
| 15 | Ferrothorn | 1 | 10.00% | 100.0% |
| 15 | Mew | 1 | 10.00% | 100.0% |
| 15 | Suicune | 1 | 10.00% | 100.0% |
| 15 | Talonflame | 1 | 10.00% | 100.0% |
| 15 | Weavile | 1 | 10.00% | 100.0% |
| 15 | Blaziken | 1 | 10.00% | 0.00% |
| 15 | Breloom | 1 | 10.00% | 0.00% |
| 15 | Chesnaught | 1 | 10.00% | 0.00% |
| 15 | Clefable | 1 | 10.00% | 0.00% |
| 15 | Excadrill | 1 | 10.00% | 0.00% |
| 15 | Gyarados | 1 | 10.00% | 0.00% |
| 15 | Hoopa-Unbound | 1 | 10.00% | 0.00% |
| 15 | Hydreigon | 1 | 10.00% | 0.00% |
| 15 | Metagross | 1 | 10.00% | 0.00% |
| 15 | Porygon2 | 1 | 10.00% | 0.00% |
| 15 | Terrakion | 1 | 10.00% | 0.00% |
| 15 | Victini | 1 | 10.00% | 0.00% |
| 15 | Virizion | 1 | 10.00% | 0.00% |
+ ---- + ------------------ + --- + ------- + ------ +
Code:
+ ---- + ------------------ + --- + ------- + ------ +
| Rank | Pokemon | Use | Usage % | Win % |
+ ---- + ------------------ + --- + ------- + ------ +
| 1 | Landorus-Therian | 7 | 70.00% | 42.86% |
| 2 | Thundurus | 6 | 60.00% | 33.33% |
| 3 | Kangaskhan | 5 | 50.00% | 60.00% |
| 4 | Aegislash | 3 | 30.00% | 33.33% |
| 4 | Amoonguss | 3 | 30.00% | 33.33% |
| 4 | Heatran | 3 | 30.00% | 33.33% |
| 7 | Talonflame | 2 | 20.00% | 100.0% |
| 7 | Charizard | 2 | 20.00% | 50.00% |
| 7 | Keldeo | 2 | 20.00% | 50.00% |
| 7 | Sylveon | 2 | 20.00% | 50.00% |
| 7 | Zapdos | 2 | 20.00% | 50.00% |
| 7 | Gardevoir | 2 | 20.00% | 0.00% |
| 7 | Kyurem-Black | 2 | 20.00% | 0.00% |
| 14 | Azumarill | 1 | 10.00% | 100.0% |
| 14 | Deoxys-Attack | 1 | 10.00% | 100.0% |
| 14 | Gengar | 1 | 10.00% | 100.0% |
| 14 | Infernape | 1 | 10.00% | 100.0% |
| 14 | Jirachi | 1 | 10.00% | 100.0% |
| 14 | Landorus | 1 | 10.00% | 100.0% |
| 14 | Metagross | 1 | 10.00% | 100.0% |
| 14 | Rhyperior | 1 | 10.00% | 100.0% |
| 14 | Rotom-Wash | 1 | 10.00% | 100.0% |
| 14 | Salamence | 1 | 10.00% | 100.0% |
| 14 | Tangrowth | 1 | 10.00% | 100.0% |
| 14 | Terrakion | 1 | 10.00% | 100.0% |
| 14 | Weavile | 1 | 10.00% | 100.0% |
| 14 | Blastoise | 1 | 10.00% | 0.00% |
| 14 | Cresselia | 1 | 10.00% | 0.00% |
| 14 | Entei | 1 | 10.00% | 0.00% |
| 14 | Gyarados | 1 | 10.00% | 0.00% |
| 14 | Rhydon | 1 | 10.00% | 0.00% |
| 14 | Venusaur | 1 | 10.00% | 0.00% |
+ ---- + ------------------ + --- + ------- + ------ +Not yet available!
Here are all teams used, the L in front of a team implies the team has lost. The W in front of a team implies the team has won:
http://puu.sh/mKnH0/56f5f60f68.txt
However, if you like visual stuff more like me, you can click on these hide tags below!
I hope you enjoy! n__n:
L |
W |
W |
L |
W |
L |
W |
L |
W |
L |
*
*Laga did not reveal his Mega in his game vs BLOOD TOTEM
W |
W |
L |
W |
L |
W |
L |
W |
L |
*Laga did not reveal his Mega in his game vs BLOOD TOTEM
L |
*W |
W |
L |
W |
L |
W |
L |
W |
L |
*Braverius did not reveal his Mega in his game vs Laga
*W |
W |
L |
W |
L |
W |
L |
W |
L |
*Braverius did not reveal his Mega in his game vs Laga
W |
L |
W |
L |
L |
W |
L |
W |
* |
* |
*TheFourthChaser disconnected from his game with Laga. There hasn't been a proper recreation yet, but these are the teams used in the match.
L |
W |
L |
L |
W |
L |
W |
* |
* |
*TheFourthChaser disconnected from his game with Laga. There hasn't been a proper recreation yet, but these are the teams used in the match.
I will keep this thread updated with more SPL games and teams used! :D
Last edited by a moderator:
WHAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAATL |/////
Stratos
Banned deucer.
- i deleted it because its useless but i still want to keep my likes -
Last edited:
https://docs.google.com/spreadsheet...Jns69WESNmfJwXidMIMtEWD1Q/edit#gid=1763006172
Doubles OU Spring Seasonal Usage Stats
kudos to Level 51 for the template and all the previous versions
everything is tip top up to date but I still have to do the visualizer and stuff
if you have any questions or if you see any errors feel free to pm me
Doubles OU Spring Seasonal Usage Stats
kudos to Level 51 for the template and all the previous versions
everything is tip top up to date but I still have to do the visualizer and stuff
if you have any questions or if you see any errors feel free to pm me
Something something statistics tl;dr. A while ago, a mysterious user by the name of leUxie happened to post a rather unfortunate post on the forums highlighting a lack of understanding about stats and useful analytical methods. Of his post, perhaps the most unfortunate part was an excerpt (which you can read here) about Weakness and Resistance analysis. Note the fact that said user failed to factor in how common attacks of various types are in his post, and also—rather humourously—the addition of a completely meaningless standard deviation value. Fortunately for leUxie (and perhaps unfortunately for everyone sane out there), leUxie is not alone in producing ridiculously uninformed posts which experience a total disconnect from reality! It's time to show off my latest stats project!
In order to determine the relative relevant bulk of most Pokemon in the format, I did a similar thing to leUxie. After determining each Pokemon's defensive type matchups, I took an average of all of them to determine a Pokemon's resistance coefficient (RC). What is different, however, is that this was a weighted average. This leads me to how the weights were determined, then.
On the page labelled type_relevancy, you can see a set of numbers corresponding to the top 90 Pokemon in the May usage stats. Essentially, each time a Pokemon has an attacking move on their set, that move's type gets 1 point; if the Pokemon gets STAB on the move, the move instead gets 1.5 points. At the end, these points are summed and the relative abundance of each type is determined. A similar analysis was done for Physical and Special moves.
One key exclusion from this process was Fake Out. Apart from Mega Kangaskhan and occasionally Mega Lopunny, Fake Out is used primarily for the flinching effect and not for damage. Thus, whenever Fake Out was used without STAB, it was considered a Status move instead. In a somewhat similar vein, Psyshock and Secret Sword were reclassified as Physical attacks since they target physical defense, while Normal-type attacks used by Sylveon or Mega Gardevoir were reclassified as Fairy-type attacks. I actually forgot to factor in the number of Sylveon running Cute Charm, oh well.
The final results can be seen in the spreadsheet; unsurprisingly, right at the top by a huge margin we have Fire-type attacks. On one hand, this could be interpreted as the result of Mono-Fire teams eating up the ladder; on the other hand, Fire-types do have a propensity to see multiple uses on the same set. It is not uncommon for Heatran (or more uncommonly, Typhlosion) to run both Eruption and Heat Wave; Charizard Y often carries Overheat alongside Heat Wave; and so forth. Notably, this is also common with Fighting-type moves (Drain Punch / Hammer Arm + Mach Punch on Conkeldurr, for example), another high contender on the list. Full details are left for the reader to peruse.
Once each Pokemon's RC is calculated, its reciprocal is then taken, such that a higher RC corresponds to better defensive capabilities. Notably, due to the RC being a reciprocal function, increases in RC are amplified greatly toward the more resistive end; having an extra resistance on a Pokemon with already good typing can mean a lot, while having an extra weakness on (say) Abomasnow isn't such a big deal since you're going to use a Focus Sash on it anyway.
After this, each Pokemon then has its uninvested (31 IV, 0 EV) stats for HP, Def and SDef calculated. Based on the earlier Physical/Special split data, the Def and SDef values are weighted and the weighted average is multiplied by the HP stat. This does in fact correspond to the damage formula, where general "bulk" is a function of the product of HP and the relevant defense stat. Finally, these bulk stats are multiplied by the RC to give an idea of the Pokemon's "relevant bulk". While these do take numerical forms, they are not intended to be used as exact comparisons, especially given the rather inaccurate and non-representative nature of the Pokemon Showdown ladder. While larger gaps are easier to trust (for example, Cresselia's score of 110442.0044 as compared to Thundurus' 60157.89705), smaller gaps should be taken with caution; a bulky Thundurus will certainly stay on the field longer than a Choice Specs Thundurus-Therian, despite the latter having a slightly higher "relevant bulk" stat due to Volt Absorb. Therefore, in general, a Pokemon having a higher "relevant bulk" than another could possibly be expected to stay on the field longer over the course of many battles, but variations in EV spreads and so forth do make a large difference.
Finally, this model is not without fault. Some things this approach does not account for:
In order to determine the relative relevant bulk of most Pokemon in the format, I did a similar thing to leUxie. After determining each Pokemon's defensive type matchups, I took an average of all of them to determine a Pokemon's resistance coefficient (RC). What is different, however, is that this was a weighted average. This leads me to how the weights were determined, then.
On the page labelled type_relevancy, you can see a set of numbers corresponding to the top 90 Pokemon in the May usage stats. Essentially, each time a Pokemon has an attacking move on their set, that move's type gets 1 point; if the Pokemon gets STAB on the move, the move instead gets 1.5 points. At the end, these points are summed and the relative abundance of each type is determined. A similar analysis was done for Physical and Special moves.
One key exclusion from this process was Fake Out. Apart from Mega Kangaskhan and occasionally Mega Lopunny, Fake Out is used primarily for the flinching effect and not for damage. Thus, whenever Fake Out was used without STAB, it was considered a Status move instead. In a somewhat similar vein, Psyshock and Secret Sword were reclassified as Physical attacks since they target physical defense, while Normal-type attacks used by Sylveon or Mega Gardevoir were reclassified as Fairy-type attacks. I actually forgot to factor in the number of Sylveon running Cute Charm, oh well.
The final results can be seen in the spreadsheet; unsurprisingly, right at the top by a huge margin we have Fire-type attacks. On one hand, this could be interpreted as the result of Mono-Fire teams eating up the ladder; on the other hand, Fire-types do have a propensity to see multiple uses on the same set. It is not uncommon for Heatran (or more uncommonly, Typhlosion) to run both Eruption and Heat Wave; Charizard Y often carries Overheat alongside Heat Wave; and so forth. Notably, this is also common with Fighting-type moves (Drain Punch / Hammer Arm + Mach Punch on Conkeldurr, for example), another high contender on the list. Full details are left for the reader to peruse.
Once each Pokemon's RC is calculated, its reciprocal is then taken, such that a higher RC corresponds to better defensive capabilities. Notably, due to the RC being a reciprocal function, increases in RC are amplified greatly toward the more resistive end; having an extra resistance on a Pokemon with already good typing can mean a lot, while having an extra weakness on (say) Abomasnow isn't such a big deal since you're going to use a Focus Sash on it anyway.
After this, each Pokemon then has its uninvested (31 IV, 0 EV) stats for HP, Def and SDef calculated. Based on the earlier Physical/Special split data, the Def and SDef values are weighted and the weighted average is multiplied by the HP stat. This does in fact correspond to the damage formula, where general "bulk" is a function of the product of HP and the relevant defense stat. Finally, these bulk stats are multiplied by the RC to give an idea of the Pokemon's "relevant bulk". While these do take numerical forms, they are not intended to be used as exact comparisons, especially given the rather inaccurate and non-representative nature of the Pokemon Showdown ladder. While larger gaps are easier to trust (for example, Cresselia's score of 110442.0044 as compared to Thundurus' 60157.89705), smaller gaps should be taken with caution; a bulky Thundurus will certainly stay on the field longer than a Choice Specs Thundurus-Therian, despite the latter having a slightly higher "relevant bulk" stat due to Volt Absorb. Therefore, in general, a Pokemon having a higher "relevant bulk" than another could possibly be expected to stay on the field longer over the course of many battles, but variations in EV spreads and so forth do make a large difference.
Finally, this model is not without fault. Some things this approach does not account for:
- Intimidate
- Weather conditions (Sand Stream Tyranitar raising SDef, Drought Charizard Y removing Water weakness)
- EV spreads
- Usefulness of the Pokemon (Mega Aggron has the highest relevant bulk but the fact that it's a Mega and that its support movepool is so shallow brings it down a few notches in terms of viability)
in terms of the raw stats, you shouldn't need to calculate anything about Pokemon's standardized bulk relative to one another. You should use X-Act's Special Tankiness and Physical Tankiness metrics for this. He was a statistical master. His metric is age-old and works well.
Neat idea trying to factor in typing as well. He never included it. As a bit of a statistician myself, I see plenty of issues with this analysis, for what it's worth, but it's at least not totally wrong like some other users' posts :^)
EDIT: Also you've gotta fix something about your formulas. Right now if you sort by anything it breaks references. Not exactly useful that way.
Last edited: