Analysis of Pokemon Usages: Which Pokemon were used by high-ranked players in May?

[X-Act]

Shoddybattle statistics provide both weighted and unweighted usages of Pokemon. Since the weighted usage of a Pokemon is simply the sum of the ladder rankings of the players using it at that time, then we can calculate the average player ladder ranking for each Pokemon as follows:

Average Player Ladder Ranking = Weighted Usage / Unweighted Usage

This can tell us which Pokemon are used by the best players.

I decided to do this for the top 250 Pokemon in the weighted list. I then found the average and standard deviation ranking for these 250 Pokemon. Those Pokemon that are above the mean+standard deviation were then considered to be used by 'expert' players. Those Pokemon that are below the mean-standard deviation were considered to be used by 'novice' players. All the others (being within one standard deviation from the mean) were deemed to be used by 'normal' players.

Here are, then, the Pokemon used by the 'expert' players, and their ranking in May's weighted list:

  #  Pokemon     Average Player Rating
--------------------------------------
 46  Wobbuffet   1357.195234
 13  Celebi      1354.885924
 20  Zapdos      1332.453283
 69  Mesprit     1330.872288
  8  Heatran     1327.69591
 43  Roserade    1324.902768
 68  Shaymin     1318.228349
 26  Suicune     1310.31188
  5  Tyranitar   1309.461581
 25  Magnezone   1307.620508
 37  Deoxys-S    1306.65134
  6  Bronzong    1299.188471
 19  Heracross   1299.035911
 22  Azelf       1293.356838
  1  Garchomp    1293.01291
 28  Jirachi     1291.013242
 14  Gliscor     1290.837871
  4  Gyarados    1290.809299
 85  Walrein     1288.95022
 41  Mamoswine   1285.562532
 12  Starmie     1276.957552
 23  Cresselia   1276.578109
 36  Tentacruel  1273.46372
 11  Metagross   1272.099143
 86  Moltres     1269.505804
 27  Machamp     1267.65269
 49  Abomasnow   1265.015125
154  Skuntank    1264.743676
 67  Uxie        1264.036991
 34  Hippowdon   1263.70761
  7  Lucario     1263.495902
 42  Dugtrio     1263.395664
 21  Vaporeon    1262.825089
 48  Spiritomb   1258.000472
  2  Gengar      1257.158098
 10  Salamence   1253.657336
  3  Blissey     1253.070887
 15  Swampert    1252.995923
139  Nidoqueen   1251.542515
 29  Breloom     1248.96794
 16  Skarmory    1248.575719
 18  Forretress  1248.102821

Now the Pokemon used by 'normal' players:

  #  Pokemon     Average Player Rating
--------------------------------------
 81  Porygon2    1242.6028
 35  Scizor      1241.728662
 96  Jumpluff    1241.412674
 54  Mismagius   1239.275428
  9  Infernape   1239.006717
 61  Slowbro     1232.74
 24  Togekiss    1231.975919
 78  Raikou      1230.705657
 33  Yanmega     1230.54624
 53  Empoleon    1230.094602
106  Kabutops    1228.758886
 32  Snorlax     1225.524651
121  Regirock    1222.068589
 39  Milotic     1220.202825
238  Phanpy      1217.5
 17  Weavile     1214.784937
 74  Hitmontop   1213.858407
 51  Kingdra     1213.728823
 70  Staraptor   1212.975024
120  Gastrodon   1212.836672
 38  Jolteon     1211.941711
102  Rotom       1211.731164
141  Torkoal     1210.938895
103  Registeel   1210.841805
 45  Donphan     1210.682514
 57  Weezing     1208.719339
 72  Cradily     1206.512776
 93  Hariyama    1204.686803
142  Linoone     1203.860382
 65  Steelix     1203.782772
 50  Gallade     1203.663296
104  Miltank     1203.607184
 58  Froslass    1202.443403
245  Tangela     1200.827586
111  Lanturn     1199.774319
 92  Azumarill   1198.445727
 63  Smeargle    1197.380235
125  Electrode   1197.096718
 94  Cloyster    1197.094995
 30  Dusknoir    1196.824439
 62  Tangrowth   1196.661039
 76  Shuckle     1195.713312
130  Poliwrath   1195.628024
210  Ditto       1192.482014
144  Omastar     1192.179707
140  Tauros      1188.171982
 59  Crobat      1185.166705
 84  Medicham    1184.204279
126  Slaking     1183.775439
175  Dodrio      1182.819588
 87  Regice      1182.719549
 88  Claydol     1180.979618
 79  Floatzel    1180.14428
 95  Honchkrow   1179.824666
171  Camerupt    1179.388753
 44  Dragonite   1178.413881
129  Qwilfish    1177.133136
 77  Drapion     1176.616187
100  Ursaring    1175.091714
162  Butterfree  1174.003552
199  Stantler    1172.043478
134  Venomoth    1171.84839
 60  Rhyperior   1171.574608
 98  Toxicroak   1171.148014
 64  Sceptile    1170.484907
161  Probopass   1169.687609
 82  Clefable    1169.234206
178  Lunatone    1167.735376
128  Cacturne    1164
146  Lopunny     1161.961741
105  Glaceon     1161.129588
132  Jynx        1159.761523
 97  Lapras      1159.489806
 75  Ludicolo    1158.325389
169  Vileplume   1158.244541
143  Exeggutor   1157.132319
170  Regigigas   1156.894737
123  Shedinja    1156.750567
133  Hitmonlee   1156.623872
168  Kecleon     1156.138593
118  Persian     1152.194703
209  Purugly     1150.609589
 83  Gardevoir   1150.272934
 40  Ninjask     1149.684258
229  Magcargo    1148.918367
127  Banette     1148.673874
147  Kangaskhan  1148.369935
190  Sableye     1147.313725
117  Drifblim    1146.327182
 73  Ambipom     1145.680093
 47  Porygon-Z   1143.738164
196  Yanma       1142.876238
206  Exploud     1142.037037
148  Articuno    1140.745124
 66  Aerodactyl  1140.279823
116  Nidoking    1139.445286
 89  Flygon      1139.444258
241  Corsola     1138.921053
151  Relicanth   1138.891365
172  Clamperl    1137.281324
202  Granbull    1137.104972
 31  Electivire  1136.432464
112  Ninetales   1135.481343
193  Arbok       1135.261062
 55  Umbreon     1134.835826
152  Rapidash    1134.068613
150  Zangoose    1130.762846
131  Hypno       1130.399414
219  Dewgong     1128.032258
 80  Venusaur    1127.707642
181  Politoed    1127.324022
101  Swellow     1127.077504
 90  Absol       1126.968314
122  Pinsir      1126.674402
176  Quagsire    1126.565432
113  Marowak     1125.890166
135  Lickilicky  1123.87714
119  Leafeon     1123.597183
197  Xatu        1122.10101
203  Parasect    1121.442623
158  Manectric   1119.110048
184  Noctowl     1118.821086
136  Scyther     1118.491753
 99  Houndoom    1116.338446
180  Gorebyss    1114.887399
221  Girafarig   1112.057377
224  Castform    1110.663793
124  Espeon      1110.605457
107  Blaziken    1110.51323
174  Bibarel     1110.314286
 52  Alakazam    1109.786343
149  Muk         1109.445302
177  Grumpig     1109.132275
240  Pelipper    1108.746835
110  Feraligatr  1107.551724
215  Stunky      1107.316547
108  Magmortar   1106.223691
 71  Arcanine    1105.421611
153  Hitmonchan  1105.255072
212  Diglett     1105.195946
198  Magby       1104.2
137  Sandslash   1103.965874
249  Huntail     1103.666667
179  Mantine     1101.063325
195  Sudowoodo   1101.032558
 56  Charizard   1100.075112
 91  Blastoise   1097.92863
192  Solrock     1097.919847
138  Slowking    1097.620083
213  Drifloon    1097.61745
243  Sunflora    1093.507246
166  Armaldo     1092.763573
239  Cherrim     1091.156627
186  Shiftry     1090.05296
183  Octillery   1088.118457
200  Glalie      1086.292308
227  Seviper     1085.508621
165  Primeape    1083.747405
188  Raticate    1079.845395
109  Typhlosion  1078.606987
115  Rampardos   1075.177207
160  Mr. Mime    1074.007825
225  Dusclops    1072.391667
145  Sharpedo    1066.864212
157  Ampharos    1066.784024
114  Torterra    1064.909866
236  Trapinch    1063.922222
218  Furret      1060.037594
216  Mightyena   1058.166667
231  Vigoroth    1056.494949
233  Bellossom   1055.739583
242  Masquerain  1055.432099
182  Dunsparce   1054.390501
208  Taillow     1052.710059

And the Pokemon used by 'novice' players:

  #  Pokemon     Average Player Rating
--------------------------------------
217  Chatot      1044.846715
237  Delibird    1038.967391
250  Delcatty    1038.482759
232  Ariados     1035.510204
204  Electabuzz  1034.502674
214  Magmar      1030.38961
223  Entei       1027.095238
156  Pikachu     1024.890625
164  Aggron      1022.822764
167  Altaria     1022.53367
234  Phione      1019.515152
159  Kingler     1018.547584
244  Bonsly      1015.054795
189  Bastiodon   1008.729452
194  Chansey     1006.545082
185  Golem       1005.626437
163  Raichu      1005.421538
226  Crawdaunt   1000.904762
207  Vespiquen    985.420765
155  Meganium     984.7003995
220  Mawile       981.9785714
205  Flareon      971.8762887
201  Beedrill     971.4953704
211  Victreebel   963.5847953
191  Wailord      948.8045603
173  Golduck      946.4063745
235  Wigglytuff   944.745098
248  Tropius      940.7941176
247  Haunter      913.6944444
187  Luxray       901.2811671
222  Ledian       894.8445946
246  Mothim       880.038961
230  Rattata      853.129771
228  Pidgeot      667.2096774

In all three lists above, the bolded Pokemon are those that are considered OU right now, while the italics ones are those that are probably going to be OU come next July 1st.

Unsurprisingly, most Pokemon that are used by 'expert' players are OU. The Pokemon being used by the highest ranked players on average, interestingly, is none other than Wobbuffet, while Electivire is the Pokemon in OU with the lowest ranked players using it. The average player ranking out of that list is 1148.53, and Electivire's player ranking is even lower than this average. So is that of Porygon-Z, and Ninjask is very near this average too. These three OU Pokemon (Electivire, Ninjask and Porygon-Z) are kinda the 'noob' OU Pokemon.

[imperfectluck]

This is a very interesting thread. Will this be repeated for June's statistics? I also find it interesting how I use the 3 highest ranked Pokemon on that list.

[the_artic_one]

154 Skuntank 1264.743676

This was the most surprising to me, what do the experts know about skuntank that the rest of us don't?

[X-Act]

Originally Posted by Fat imperfectluck

This is a very interesting thread. Will this be repeated for June's statistics? I also find it interesting how I use the 3 highest ranked Pokemon on that list.

I'll do it again for June's statistics if you wish.

[animenagai]

please do this again x-act, i love it! i used wobb extensively in may, so i must be good :D. this should add to the 'wobb is uber' arguments. keep in mind that lesser used pokes can look real good just because a few really good players use them. no, this should not make skuntank OU.

since i'm already at it, might as well post some thoughts on why these pokes are ranked where they are. there are lots of pokemon that i find hard to fit on a team despite their high rankings. these are:

1. infernape. LO + sandstorm just proved too much to me. don't forget SR damage either. too much damage on a fragile poke.

2. scizor. it's just too easily walled. i can see potential with niche roles but flying types just love facing scizor, especially gyara and mence due to intimidate.

3. yanmega. 50% SR damage is just no good. so many people use it with LO too. it can wear down really easily especially when the current metagame is filled with bulky sweepers.

4. weavile. bronzongs are running all over the place and gengars - something weaviles were known to trap kill, now commonly run scarves.

5. snorlax. it's just hard using normal type physical threats. skarm, zong, dusknoir... you get the picture. there are better rest talkers.

[Trust]

Some surprises:
- Mesprit: it's quite useful, but I don't find a way to use it correctly. However some people have found its usefulness.
- Shaymin: it's barely used, but who use it are quite strong. I only see Specs set, which have a limited movepool, so I don't understand why it's so used.
- Walrein: with Walrein so high, we can find that Hail teams are used by strong people.
- Skuntank: I don't understand this, sorry.
- Nidoqueen: I've used it with success, because it's a good Heracross and SDLucario counter, can lay down Toxic Spikes, Stealth Rock, absorb that Toxic Spikes, and and with Roar it can PHaze quite well.
- Porygon2 > PorygonZ: do you understand this?.

Anyway, what are we missing about Mesprit, Shaymin and Skuntank?

[Blue Kirby]

I'm not surprised by Mesprit at all - I use it on a lot of my teams with plenty of success. Skuntank sticks out like a sore thumb to me on the list - I can "justify" the other highly ranked Pokemon; each has a niche it fills in some way or another. I'll definitely be looking into this.

Excellent work, as always, X-Act. Much appreciated. :)

[Deco]

Nice work as always, X-Act. ^^

I'm not surprised about most things, really - I've always considered Ninjask and Electivire n00b-OU Pokemon, as most people who still think those are good Pokemon are n00bs anyway.

Also, Wobbuffet at first place might be a good argument for it being uber... one could conclude that using Wobbuffet makes your rating increase. x3

I'm kinda wtf'ing at Mesprit being so high. o.o;

[CardsOfTheHeart]

Mesprit has the ability to counter Gyarados, Garchomp, and Infernape, among other things. Should its high rating really be such a shock? (Okay, higher than Heatran is a little bit of a surprise...)

The real WTF is, of course, Skuntank. Heh, must be for the WTF factor. lol

[EvilBob]

I'm guessing the deal with Skuntank is that it's pretty low on the average usage list, so being on just a couple teams that ended up moderately high in ranking would bump it up pretty high.

[kd24]

Yeah. But still, Im always glad to see Skuntank ranked above Lucario.

[X-Act]

Well, maybe there was a high-ranking player in May who was experimenting with Skuntank.

EDIT: These were the Pokemon used by 'experts' in April:

  #  Pokemon     Average Player Rating
--------------------------------------
191  Sunkern     1482.336957
 43  Wobbuffet   1351.264452
 15  Celebi      1339.323781
 73  Mesprit     1330.504567
 47  Deoxys-S    1329.663772
 22  Zapdos      1319.160436
  5  Tyranitar   1318.568092
  8  Heatran     1313.558359
 26  Jirachi     1311.774371
 14  Gliscor     1300.457125
 82  Shaymin     1296.629252
  1  Garchomp    1291.060546
 20  Azelf       1289.938249
  4  Gyarados    1288.767375
 44  Dugtrio     1287.248012
 50  Roserade    1286.397905
  7  Bronzong    1285.640865
 32  Mamoswine   1285.632671
 21  Cresselia   1284.963119
 33  Magnezone   1282.471658
 18  Heracross   1277.668828
  9  Starmie     1277.544353
  6  Lucario     1276.314371
 31  Machamp     1270.386419
 27  Hippowdon   1267.985191
 40  Suicune     1267.741505
 96  Uxie        1265.273377
 88  Porygon2    1264.105828
 16  Swampert    1262.550627
 25  Breloom     1260.624922
 51  Abomasnow   1258.722864
 41  Tentacruel  1255.958439
 56  Mismagius   1254.569389
 13  Skarmory    1254.417011
 75  Cradily     1251.752231
 12  Metagross   1251.315439
 24  Snorlax     1250.697852
 17  Forretress  1249.755686
103  Kabutops    1248.134581
  2  Gengar      1247.068385

There seems to have been a relatively-high ranking Sunkern gimmick user in April.

[Amelia]

... sunkern...
Now that's even more weird... but nice job X-Act!

[TheMaskedNitpicker]

Thanks for putting this together, X-Act. It's definitely some very interesting info.

Originally Posted by Fat Deco

Also, Wobbuffet at first place might be a good argument for it being uber... one could conclude that using Wobbuffet makes your rating increase.

Nope. Correlation does not imply causality. Any number of other explanations are possible. For instance, it could be that good players tend to be the ones that ignore the 'code of honor' surrounding Wobbuffet and therefore make up a very high percentage of its users. It could also mean that using Wobbuffet successfully takes a great amount of skill and that only expert players can use it to good effect. Seems to me that this last conclusion would be a good argument for it staying in OU, since one of the popular reasons for banning it is that it 'removes skill from the game'.

Originally Posted by Fat EvilBob

I'm guessing the deal with Skuntank is that it's pretty low on the average usage list, so being on just a couple teams that ended up moderately high in ranking would bump it up pretty high.

Good thinking. X-Act, is it possible that you could include the number of unweighted usages on this chart as well? This could give us a bit more info on the 'fluke' factor for each of these Pokémon.

EDIT: I have a question. When 'weighted statistics' are calculated, what's the exact algorithm? Is it the ranking of the player when they start the match, after they finish the match, or does Shoddy keep track of the number of uses for each Pokémon on an account over the month and then multiply by the month-end ranking?

[Jumpman16**]

I love this a lot, thanks X-Act. I will remind you and everyone else though (because it hasn't been posted yet), that the preponderance of alternate accounts is going to throw these numbers off an awful lot. You almost need look no farther than the top two players on the ladder right now to realize this: they are both "imperfectluck" for whatever reason.

His analajim account is #6 as well, and for a player who's voiced his goal of whoring Wobbuffet until it gets banned, you can start to realize one of the main reasons Wobbuffet is so high on this index. I honestly don't know whatever alts he has or other players have, but until we're able to pin this down, we unfortunately cannot place much stock in these stats.

[Imran]

Hmmmm, I'm not sure about this. The first 5 pokemon on the list used by expert Players are Legendaries, with the exception of the Wobb, which has recently been moved down from Ubers. This seems to go against the old phrase "Those who battle without legends, don't have the skills to deal with them." Seeing as the "Experts" use the legends, then doesn't that seem to suggest that they rely on their legends, to be "experts."

The other explination is that... LOL I'm ill and i can't remember what i was going to say. Either way there was an alternative explaination. Anyway the conclusion of this small piee of writing is that bit about the legends earlier!

[EvilBob]

Originally Posted by Fat TheMaskedNitpicker

Good thinking. X-Act, is it possible that you could include the number of unweighted usages on this chart as well? This could give us a bit more info on the 'fluke' factor for each of these Pokémon.

Although would be nice information to see, he does list Skuntank's position on the usage chart, which is only 154 (163 on the unweighted list), which is definitely low enough for a couple high ladder players to bring up the usage ranking.

[X-Act]

Originally Posted by Fat TheMaskedNitpicker

Good thinking. X-Act, is it possible that you could include the number of unweighted usages on this chart as well? This could give us a bit more info on the 'fluke' factor for each of these Pokémon.

I think listing the fact that it's only the 154th used Pokemon is good enough. :) If you're curious, it was used 593 times in May.

EDIT: I have a question. When 'weighted statistics' are calculated, what's the exact algorithm? Is it the ranking of the player when they start the match, after they finish the match, or does Shoddy keep track of the number of uses for each Pokémon on an account over the month and then multiply by the month-end ranking?

From the Shoddybattle site:

Each usage scores the pokemon a number of points equal to its trainer's conservative estimated rating at the point when this list was compiled...

I'd guess that means that it counted the players' ratings as they were in June 1st.

[TheMaskedNitpicker]

Originally Posted by Fat X-Act

I think listing the fact that it's only the 154th used Pokemon is good enough. :) If you're curious, it was used 593 times in May.

Thanks!

From the Shoddybattle site:

Each usage scores the pokemon a number of points equal to its trainer's conservative estimated rating at the point when this list was compiled...

I'd guess that means that it counted the players' ratings as they were in June 1st.

Interesting. That means that if a player's skill improved a lot over the course of a month, their 'noobish' teams from the beginning of the month have just as much effect on the weighted totals as their 'expert' teams near the end of the month. Hardly an ideal system, I'd say.

Originally Posted by Fat Jumpman16

His analajim account is #6 as well, and for a player who's voiced his goal of whoring Wobbuffet until it gets banned, you can start to realize one of the main reasons Wobbuffet is so high on this index. I honestly don't know whatever alts he has or other players have, but until we're able to pin this down, we unfortunately cannot place much stock in these stats.

Maybe I'm misunderstanding how these statistics are calculated, but I don't think that a player having multiple high-ranking accounts makes these stats any different than if that player had a single high-ranking account. For weighted usage, it's just ranking multiplied by number of usages. It shouldn't matter much if those usages are split among several accounts as long as all those accounts have about the same ranking at the end of the month.

[Taylor]

Great work, X-Act.

I'm not too suprised that Wobbuffet is used the majority of the time by high-ranked players on the ladder. As such, this is easy to acknowledge once you realise that ipl has two accounts on the top ten of the Shoddy Battle ladder. On the other hand, you see, many lower-tiered players don't tend to use Wobbuffet due to their misundertanding of the Pokemon in general - believing it's not that great - or the fact that they won't use it because they simply do not like it - hatred fumed through Wobbuffet's "uberness". Don't get me wrong, though, as there are many higher-tiered players who simply do not use Wobbuffet either.

[Dr.Traveler]

Thanks for the work.

Interesting to find that of the 6 on my current team, 5 are in the normal category and 1 in the Expert. I definitely should play more.

[Great Sage]

I wasn't particularly surprised by any of this, but of course, awesome work, X-Act. Electivire as the bottom OU sounds pretty much right, though its usage seems to be dropping; I only saw 1 of them so far this month.

[BlueCookies]

I'll take the credit for the Walrein usage, as for most of May I used a hail team to great affect. It surprisingly dominated nearly everyone I played against, so I guess that's why he was so high.

Who was the one who used Sunkern again, I remember a few people were talking about it in the main chat or somewhere, I don't remember. Something makes me think it was KinglerDude, but I'm not sure.

So let me get this straight, if someone ranked at the top(IPL) used Magicarp/Sunkern/Metapod/Weedle/whatever for only one match, they would all be at the top of the list because they were probably only used by him the entire month? So there would be no way to find an average becuse only oneplayer used them?

[Jumpman16**]

Originally Posted by Fat TheMaskedNitpicker

Maybe I'm misunderstanding how these statistics are calculated, but I don't think that a player having multiple high-ranking accounts makes these stats any different than if that player had a single high-ranking account. For weighted usage, it's just ranking multiplied by number of usages. It shouldn't matter much if those usages are split among several accounts as long as all those accounts have about the same ranking at the end of the month.

The way I look at it, five good people set on whoring the shit out of Wobby could have made accounts that comprise 50% of the top-100 ranked players (I'm exaggerating for a reason). To me, this indicates more that these five players are good/have a lot of time on their hands than that Wobbuffet is the pokemon most used by the players on the ladder.

To be totally semantic and nitpicky in order to be as clear as possible, the title of this thread would be more accurate if "players" were replaced with "accounts". I may not be thinking about this the right way either, though, and I'd appreciate it if X-Act could weigh in.

[TheMaskedNitpicker]

Originally Posted by Fat Jumpman16

The way I look at it, five good people set on whoring the shit out of Wobby could have made accounts that comprise 50% of the top-100 ranked players (I'm exaggerating for a reason). To me, this indicates more that these five players are good/have a lot of time on their hands than that Wobbuffet is the pokemon most used by the players on the ladder.

To be totally semantic and nitpicky in order to be as clear as possible, the title of this thread would be more accurate if "players" were replaced with "accounts". I may not be thinking about this the right way either, though, and I'd appreciate it if X-Act could weigh in.

Well, let's say that a great player plays 5000 matches all on one account. Another great player has five accounts and plays 1000 matches on each. Using these statistics, I don't think the second player's usages count any more than the first player's. I thought is was based on total usages, not percentage of usage per account.

EDIT: Wait, I think I see what you're saying. The more accounts there are, the lower everyone else's ranking is. Is that right? How is the ranking actually calculated? I suppose I should investigate before making further hypotheses.

If players are actually whoring Wobbuffet in an effort to get it banned, that's very poor form. That doesn't show that Wobbuffet's powerful. It just shows that good players are using it in order to artificially raise its usage. Not cool.