[Competitive] Tournament Usage Stats [Update Post #30]

[Darknut07]

but neither does an ad ignorantium prove that the status quo is accurate. As it stands you've done nothing to back up your claims that Tyranitar is the better setter, and Lavos provides some good points.
[/size]

Hold on, when someone makes a claim that a pokemon that has been for (pokemon) generations been considered one of the best in the game is inferior, surely the burden of proof falls on them to provide evidence. For example, and at an extreme, I can claim without proof that Scizor is superior to Scyther. Still, I did not make this claim or any about Tyranitar, but simply said these stats offer no proof to the superiority of one or the other. If we were to go by the logic employed by the poster who claimed otherwise, we would have to concede that Amoonguss is more useful than a great many pokemon. Rather, neither Hippo nor Amoonguss is as useful as the thread suggests and their inflated usage statistics are likely a by-product of a small sample size.

 

[Pikachuu]

considering the current tournament metagame trends, i actually would go as far as to say that hippowdon is a superior sand inducer to tyranitar. it's far bulkier, sets rocks easier, has access to instant 50% recovery, phazing, and a powerful stab earthquake. tyranitar's only real advantages are a decent special movepool and access to pursuit. considering that ttar loses to all the top 5 tournament ou pokemon, whereas hippo loses to only 2 of those 5 (and also isn't trapped and killed by dugtrio), i think it's fair to say that, given the amount of data we have on the tournament metagame, hippowdon outclasses tyranitar.

Tyranitar is a lot better than Hippowdon imo. Hippowdon gets completely walled by Xatu on sun teams and falls to offensive Politoed and it won't enjoy taking a Fire Blast from Ninetales and bad match up against Abomasnow. Also if you really need a physical wall you can always pair up Tyranitar with something like Slowbro or Gliscor. Hippowdon is also Taunt bait. Hippowdown also has it's own problems against trappers like Choice Specs Gothitelle not just Tyranitar.

 

[kd24]

ill try to update these by tonight

 

[Lavos]

Tyranitar is a lot better than Hippowdon imo. Hippowdon gets completely walled by Xatu on sun teams and falls to offensive Politoed and it won't enjoy taking a Fire Blast from Ninetales and bad match up against Abomasnow. Also if you really need a physical wall you can always pair up Tyranitar with something like Slowbro or Gliscor. Hippowdon is also Taunt bait. Hippowdown also has it's own problems against trappers like Choice Specs Gothitelle not just Tyranitar.

i mean you can spew all that out there and on paper it may or may not be true but the fact of the matter is that over the course of thousands of bw2 ou battles i've come to realize it's pretty clear that more often than not hippowdon is going to be the more valuable member of the team than tyranitar. maybe i didn't elaborate on my points enough but my competitive experience tells me hippo outclasses ttar most of the time. ttar has its uses and it's a great pokemon in my opinion, nothing else does quite as great of a job pursuit trapping, but for the current metagame, hippowdon's a better fit.

also thanks for continuing to do this kd24, we appreciate it

 

[kd24]

There was an error in the last stats that truncated Gengar off. It's been fixed in the total stat update that is now complete. A few changes in usage:

- Tiebreakers aren't going to be used as of now, Pokemon will continue to separate into their own usage stats every week.
- Because we had a top 24 and a top 16 cut, the number of Mons will only be listed in the specific usage rankings

Specific Usage / Raw Data: 2 teams unrecorded, statistics pulled out of 46 teams. This was for today's tour.

- I am showing this data for any stat junkies to try and find total usage errors in the detailed stats. I hope I didn't fuck anything up but if I did, its the percentages that change, not the actual rankings.

Jirachi: 11/24, 4/10, 2/6, 3/6
Rotom-W: 7/24, 4/10, 1/6, 2/6
Tyranitar: 8/24, 3/10, 1/6, 1/6
Scizor: 7/24, 2/10, 2/6, 1/6
Landorus-T: 6/24, 1/10, 2/6, 3/6
Politoed: 4/24, 2/10, 3/6, 3/6
Latias: 7/24, 3/10, 1/6, 0/6
Landorus: 6/24, 4/10, 1/6, 0/3
Garchomp: 7/24, 2/10, 1/6, 0/6
Keldeo: 5/24, 1/10, 2/6, 2/6
Ferrothorn: 4/24, 3/10, 1/6, 2/6
Forretress: 6/24, 3/10, 0/6, 0/6
Starmie: 4/24, 1/10, 1/6, 3/6
Thunderus-T: 3/24, 1/10, 2/6, 3/6
Kyurem-B: 2/24, 2/10, 2/6, 3/6
Alakazam: 2/24, 2/10, 2/6, 2/6
Gengar: 2/24, 2/10, 1/6, 2/6
Terrakion: 4/24, 1/10, 1/6, 0/6
Latios: 4/24, 1/10, 1/6, 0/6
Celebi: 3/24, 0/10, 2/6, 1/6
Lucario: 2/24, 1/10, 1/6, 2/6
Ninetales: 3/24, 2/10, 0/6, 0/6
Heatran: 3/24, 1/10, 0/6, 1/6
Skarmory: 1/24, 1/10, 1/6, 2/6
Venusuar: 2/24, 2/10, 0/6, 0/6
Dugtrio: 2/24, 2/10, 0/6, 0/6
Breloom: 3/24, 0/10, 0/6, 0/6
Tentacruel: 2/24, 1/10, 1/6, 0/6
Amoongus: 2/24, 1/10, 0/6, 0/6
Chandelure: 2/24, 1/10, 0/6, 0/6
Reuniclus: 2/24, 1/10, 0/6, 0/6
Gyarados: 2/24, 1/10, 0/6, 0/6
Hippowdon: 2/24, 1/10, 0/6, 0/6
Dragonite: 1/24, 1/10, 1/6, 0/6
Magnezone: 1/24, 1/10, 1/6, 0/6
Wobbuffet: 2/24, 0/10, 0/6, 0/6
Cresselia: 1/24, 1/10, 0/6, 0/6
Zapdos: 1/24, 1/10, 0/6, 0/6
Volcarona: 1/24, 0/10, 0/6, 0/6
Espeon: 1/24, 0/10, 0/6, 0/6
Salamence: 1/24, 0/10, 0/6, 0/6
Gorebyss: 1/24, 0/10, 0/6, 0/6
Abomasnow: 1/24, 0/10, 0/6, 0/6
Liligant: 1/24, 0/10, 0/6, 0/6
Chansey: 1/24, 0/10, 0/6, 0/6
Gliscor: 0/24, 0/10, 1/6, 0/6

These stats are then added to yesterday's statistics. The new information becomes our detailed stats. The personal Pokemon stats will be updated at a later date (I'll have 3 weeks to actually make some cool stats maybe?):

Detailed Stats

Non-Weather Usage: 26.33%
Rain Usage: 28.95%
Sun Usage: 10.52%
Sandstorm Usage: 32.89%
Hail Usage: 1.31%

1.] Jirachi: --------- 44.74%
2.] Landorus-T: ---- 31.58%
3.] Politoed: -------- 28.95%
4.] Keldeo: --------- 27.63%
5.] Scizor: ---------- 27.63%
6.] Rotom-W: ------- 26.32%
7.] Starmie: --------- 22.37%
8.] Tyranitar: ------- 22.37%
9.] Landorus: ------- 19.74%
10.] Garchomp: ------ 19.74%
11.] Thundurus-T: --- 17.11%
12.] Latias: ---------- 17.11%
13.] Skarmory: ------- 17.11%
14.] Forretress: ----- 14.47%
15.] Kyurem-B: ------ 14.47%
16.] Ferrothorn: ------ 15.79%
17] Latios: ---------- 14.47%
18.] Heatran: -------- 13.16%
19.] Gengar: ---------- 13.16%
20.] Dragonite: ------- 11.84%
21.] Celebi: ---------- 11.84%
22.] Ninetales: ------- 10.53%
23.] Breloom: -------- 10.53%
24.] Hippowdon: ----- 10.53%
25.] Alakazam: ------- 10.53%
26.] Amoongus: ------ 9.21%
27.] Venusaur: ------- 9.21%
28.] Terrakion: ------- 9.21%
29.] Dugtrio: --------- 9.21%
30.] Gyarados: ------- 7.89%
31.] Lucario: ---------- 7.89%
32.] Chansey: -------- 6.58%
33.] Tentacruel: ------ 6.58%
34.] Donphan: -------- 5.26%
35.] Mamoswine: ----- 5.26%
36.] Weavile: -------- 5.26%
37.] Xatu: ----------- 3.95%
38.] Chandelure: ------ 3.95%
39.] Reuniclus: -------- 3.95%
40.] Magnezone: ------ 3.95%
41.] Gliscor: ---------- 2.63%
42.] Jellicent: ------- 2.63%
43.] Zapdos: --------- 2.63%
44.] Cresselia: -------- 2.63%
45.] Wobbuffet: ------ 2.63%
46.] Salamence: ------ 1.32%
47.] Liligant: --------- 1.32%
48.] Abomasnow: ----- 1.32%
49.] Sableye: -------- 1.32%
50.] Darmanitan: ----- 1.32%
51.] Tornadus: ------- 1.32%
52.] Gorebyss: -------- 1.32%
53.] Espeon: ---------- 1.32%
54.] Volcarona: ------- 1.32%

I'm going to be honest to the public. This is pretty physically draining for one person to do. I'm looking for more volunteers, people who I can quickly train on how to properly record the usage, as well as people who can help with detailed statistics. Thanks to Harsha and PKGaming for helping record some stats, although there were a few issues with them that I have now fixed. So if you think you'd be interested in helping, shoot me a PM with what you can or would be willing to actually do. I'm also looking for ways to better collect full stats more quickly, and an automated system for tallying, rather than me doing it completely by hand, where errors can occur.

Onto the stats, the Tyranitar vs Hippo debate comes to an end? I actually like Hippowdon a lot more, especially because the top 5 stayed almost exactly the same (Keldeo and Politoed flip-flopped). I at least feel glad that my Tyranitar prediction sort of came through. Note though that Landorus-I usage also went up dramatically in today's games. CONNECTION?

Forretress is actually more common than Skarmory in a general tournament environment. What is inflating Skarmory's usage is its same team winning percentage. Both finalists of Friday's tour were using Skarmory, and the winner of today's tour used it as well. Forretress was way more common in early rounds but also got knocked out in them, eliminating his late usage. Skarmory is the shit though, and the statistics show that using Skarmory = winning basically 100% of your battles. Yeah, that's pretty good. Things have begun to settle down, the biggest surprises being no one uses Salamence. Its sole usage today belonged to a SmashPass team.

Rotom-W shits on everything, you should be using it. Landorus-I cracked top 10 with an impressive performance today. Sand and Rain are neck and neck at the moment, weatherless is being represented pretty accurately in my opinion.

I'm really tired, so I don't think I'll comment any further, I'll let you guys do that. If you really like this concept of gathering tour statistics, be sure to say so, so we can try to continue progress with this down the line. I find these 100x more interesting in ladder stats and it really tracks what the meta is really like.

Jirachi iz UBERS :D

 

[PK Gaming]

Cutsap Skarmory is going to end up on everyones offensive team from now on.

ShakeItUp jokingly mentioned that it's even better than Deoxys-D, but honestly it's not that far behind it. It's really damn good at getting up hazards haha. Speaking of SIU, his Kyurem-B did massive work in today's tour. I don't think it would be unreasonable to say that its really one of the top threats in the tier. That combination of power, and ability to setup on really important Pokemon (Rotom-W / Starmie) makes it invaluable.

Landorus-T was freaking everywhere. It just goes to show that an easy SR and a pocket intimidate is extremely useful.

I'm really tired, so I don't think I'll comment any further, I'll let you guys do that. If you really like this concept of gathering tour statistics, be sure to say so, so we can try to continue progress with this down the line. I find these 100x more interesting in ladder stats and it really tracks what the meta is really like.

Agreed. Tournament stats are a norm in pretty much any competitive environment worth their salt, so I think that once we get a solid method at gathering tournament stats, we should base our tiers around tournament stats instead of ladder stats.

 

[dcae]

Its incredible how much the total use of Landorus. Both variants are being in total used on about half of the teams, which is insane when you think about it. On another note, I predicted along the lines of kd24 and saw Ttar going up top, but Jirachi is so far pulling it with some very high usage. I was surprised in the first set to see Terrakion lower than Donphan, but things seem to have righted themselves. I wholly enjoy these statistics, and I hope they can be continued.

 

[Stratos]

we should base our tiers around tournament stats instead of ladder stats.

I feel like this idea is probably based off of a misunderstanding of the purpose of tiers. It's not to tell you what's good or bad in a meta, it's to sift out Pokemon of similar power levels so that the tier below it is playable, and I really think that ladder stats do a better job of that than tourney stats. The sample size on tourneys is far too small and inconsistent (what do we do when smogon tour stops? there's no big OU tourneys for a while past that). These usage stats are infinitely better for reference, I agree, but there's no way with such a small population we could ever hope to capture every Pokemon that should be OU as being OU—something that ladder stats do a fairly strong job at currently, even if we do end up snatching up Pokemon that should be of a lower tier unfairly sometimes. I think it's a price I'd be willing to pay to not, for example, have UU Mence

Then again, this may be because I'm looking at it in the context of one add/drop at a time. If we implemented this idea swiftly and unilaterally once we had enough data points, it would surely completely upheave the tiers as we know them. but aside from the bitchers and moaners at GameFAQs, would we be able to have equally playable, more balanced, better tiers as the end result? In general, it comes down to whether you think tiers should be decided by what's overall, in general, good, or what's good for the meta as it stands. Ladder stats capture what's overall good, since you're playing for the long term, but tournament stats catch what's more effective in the meta. It'd be an interesting experiment to see if a tier of 'good-but-misfit' Pokemon like a shift to tourney stats would produce would be better than a tier of 'plain weaker' Pokemon like we currently have (since all the misfits get banned to BL)

But honestly I have my money on that ladder stats would be better for tiering

 

[PK Gaming]

I think our tiers should accurately represent the highest level of play in Pokemon.

What's "good for the meta" is irrelevant.

EDIT: It's probably not a feasible idea though. There aren't enough tournament matches for the non-OU tiers to support tier based tournament stats. Oh well...

 

[Soul Fly]

So Custap Skarmory is no longer that awesome gimmick?

shit... I had a fun time surprising people with it.

 

[Lork]

because everyone misinterpreted:

it's interesting that hippowdon is gaining momentum as the superior sandstream setter.

aka those stats were an example of a notion which is becoming more popular. not that those stats absolutely mean hippowdon is superior.

anyway thanks for the new bout of stats

 

[Harsha]

People are complaining about the small sample size, but this will obviously become a non-issue once we continue working on the project. Inflation will be diluted once we can get a larger sample size to look at. That being said, I think these stats do reflect the metagame quite accurately. Looking at Jirachi's usage, I would go as far as to call it the most diverse Pokemon in the metagame – it works as a special wall or a great SubCM user on rain offense, and it can work as a special wall, Choice Scarf user, or even an Expert Belt bluffer on pretty much any other team. I've even seen SubCM sets with Thunderbolt on sand, so it's pretty unpredictable, and you definitely have to tread cautiously around it when you see it in team preview.

Landorus-T at second is a bit of a surprise, but I've known for a while that Landorus-T would be fairly high up in the standings in tournaments, as I've seen it used a lot more in a more competitive environment. Like PK Gaming said, it works as a great defensive pivot with Stealth Rock, it has a very solid typing, and it can dole out a lot of damage with little investment (the standard 200 HP / 64 Atk / 244 Def spread with an Adamant nature is still extremely powerful!). It can run a nasty double dance set too, though it's a bit easier to know when to expect double dance on Landorus-T because another Pokemon will have to carry Stealth Rock.

I, like kd24, am pretty surprised to see Tyranitar's usage so low. I'm not saying it's the best sandstorm inducer or the best weather starter, but its benefits are so expansive that you'd expect at least a bit more Tyranitar usage. For one, Tyranitar's Pursuit support helps those heavy offense teams (Tyranitar / Terrakion / Keldeo / Scizor / Landorus / Latios, a pretty formulaic build) get rid of opposing Latios, Latias, Jellicent, Celebi, and many other Pokemon. This is very useful for late-game Keldeo sweeps, or even late-game Landorus sweeps if you're not using U-turn Landorus. I've also seen it on some sand stall as a Choice Scarf user, but that's to a lesser extent. Even with its weakness to Water-type moves, Tyranitar helps a lot against rain teams because it has weather control and Pursuit-trapping capabilities to get rid of common problematic Pokemon on rain teams.

I predicted Celebi's usage to be in the top 10, but I was pretty far off. Celebi is a great pivot, and with Baton Pass, it can't be trapped by opposing Choice Band Tyranitar. It can provide a lot of support options with access to Stealth Rock, Thunder Wave, Perish Song, and so on and so forth, which means it fits really well on anything, be it rain offense, sand offense, sand stall, or plain weatherless. I definitely expect its usage to go up after a few more Tours.

Trick Room Reuniclus is the same Pokemon as before, but I think it's starting to pick up a lot more usage. It sets up relatively easily in the late-game after a Choice Scarf Keldeo uses Surf or something similar happens, and once it's out, it can OHKO everything on weakened offensive teams. It's always been a nuisance for stall to deal with due to not taking passive damage, though Calm Mind is obviously better against stall. Either way, it wrecks most forms of stall once Tyranitar is taken care of. I don't necessarily expect its usage to spike because it's not a top 10 kind of Pokemon, but it definitely will get more popular as an anti-metagame option once Tour progresses.

I'll just echo what everyone else has said about Cutsap Skarmory. Ever since I saw Cutsap Forretress and Skarmory in the Ubers metagame, I considered trying them out in OU, and they're extremely useful. Cutsap Forretress can spinblock when it gets down to low health by using a fast Explosion, and it has access to every hazard, so you can always get at least two layers down before getting KOed. Cutsap Skarmory is very useful as well, and it does pretty much the same thing, though it has better defenses and more Speed to work with, while it doesn't have access to Explosion. Both of these Pokemon are extremely easy to spot from team preview, but yesterday ShakeItUp was still able to get at least three layers of hazards up every time I watched him play (he eventually went on to win the Tour). These are pretty good lead options for heavy offense teams that don't want to rely on Smeargle.

Finally, everyone should thank kd24 for all the effort he puts into these stats. I did a very small portion of the work yesterday and I already found it exhausting – I can't imagine having to compile all of these. I definitely appreciate the initiative he took.

 

[alkinesthetase]

commenting on the technical side here, are you doing all these calculations by hand kd? o_O busy for the next few days but nevertheless talk to me on IRC or something, i can't help with actually collecting teams (clueless in tour @_@) but i can help you automate the number crunching

 

[HabibsHotDogs]

You can't use tournament stats as official tiering stats... Its because the whole point of tiering usage stats is to gauge how likely it is to see certain threats and hence how effectively you should prepare for them on your team.

And well, the way to find out how common threats are in OU, is to base usage stats of OU, not tournaments... Tournament stats can be used to gauge the frequency of threats in tournaments.

Just ask Aldaron I'm sure he'll agree with me.

 

[Chou Toshio]

@Tobes-- anyone can tell that this took extensive effort to compile-- how is that at all relevant?

Even if it took a lot of effort, that doesn't change the fact that this little data has almost no prediction power, explanatory power, and leaves no room for actual statistical analysis. Touting around this kind of information is simply misleading to the scientifically unsophisticated reader. If that isn't an important point for this thread then smogon's gone off the deep end.

Everything I said in my post was true-- maybe I should have sugar coated it to make it nicer?

@Harsha-- by the time you get a new tournament to add more data, the meta game will have changed-- the older data will become irrelevant. Furthermore the short term build nature of tourney teams means this churning effect will be even worst. Finally, you still don't have a define able population. Statistically, there are so many problems it's not funny


There's no way to be 'surprised' or 'impressed' by any placement on the list since with so little data, everything falls into the predictable deviations-- for a population defined to the usage of a single tourney that's already over.

 

[MMII]

I want to comment a bit on the first day's statistics (I had meant to do the day they came out) just because I find it fascinating the way Peje used the same (more or less I assume) Full Stall team since the top 16 and nearly won every match.

As a Stall team, there are some interesting picks that can be seen just from the team preview. Having sand as a weather is pretty typical for Stall in this metagame but using Chansey over Blissey on it stands out. I'm not sure if this choice was made to handle Landorus-I better or because Peje expected to see heavy Rain usage (following from pre-top 16 battles?) or simply because he just prefers Chansey regardless of the weather. (in which case, I agree wholeheartedly) Another interesting choice was Amoongus over the generally more popular Celebi. This preference means he loses out on a counter to the recently hyped Lando-I in favor of a check that is non-pursuitable (at least, not like Celebi) and hard-counters every Keldeo variant including the rising HP Bug (plus Regenerator is just such a nifty ability and it could absorb TSpikes for Chansey). This was a strong metagame choice (perhaps influenced from earlier battles as well?) seeing as it could come in on Keldeo and many other Rain threats with ease and not have to waste a turn recovering off the damage. The final "oddity" that hops out in Peje's team is the lack of spin blocker. With the massive amounts of offensive threats, it is understandable why he decided to pass up on a ghost type in order to cover threats a Jellicent (or other lesser used defensive ghosts) couldn't cover (which isn't that unusual nowadays but it's certainly a change brought by BW2). However, having a solid Spikes setter in Skarmory as well as whoever he stuck SR on (all three options are capable users) helped remedy this by replacing the lost hazards easily enough (although it looks like the team would struggle in wearing down a defensive Starmie). That's all I can tell from just having a list of members, who knows what other surprises lay in the actual sets.

When matched up the metagame projected by kd24's stats, it appears that this Stall team was a strong anti-metagame pick. As already mentioned, the popularity of Rain and Keldeo made Amoongus an invaulable member as it strong switch-in against these teams. Jirachi, who was number one in this tournament, didn't have much room to breath as Hippowdon and Landorus-T could threaten it with a strong EQ , defensive(?) Starmie and Skarmory feared nothing from defensive sets (cept U-Turn for Mie) and could respond with Spin/Spikes, and even Chansey and Amoongus could play their part against certain CM variants by wearing it down or crippling it with SToss/Spore. Landorus-T and Scizor were going to have a hard time getting past the physically bulky Scizor, Landorus and possibly even Hippowdon. Skarmory wouldn't be able to keep his Spikes up long with Starmie around to Spin and threaten it out with Scald. Dragonite had to deal with Lando+Hippo+Skarm so it could just about never sweep. In fact, the powerful defensive backbone this team had on both spectrums meant it would take a lot to pull of a sweep with anything. Ferrothorn looks like to one of the important weakpoints of the team being able to constantly pester the team with Spikes/Seeds but that fortunately saw very little usage. Overall, this was definitely a nifty team and an excellent choice for the metagame at that time.

Aside from the Stall team, it looked like somebody used the same Sun team consisting of Tales/Xatu/Trio/Saur and two other members from the top 16 all the way to top 4. I'm personally interested in hearing about the growing popularity of Custap Skarm/Forry. The lack of Taunters makes them a pretty strong choice at the momment. Anyways, thanks for these cool stats kd24.

 

[Lavos]

that's user P E L E J A G A R T O you're thinking of, not masterclass. he beat me with that sand stall team in semis, going on to lose in the finals.

 

[GatoDelFuego]

What Lavos said

I'm personally interested in hearing about the growing popularity of Custap Skarm/Forry. The lack of Taunters makes them a pretty strong choice at the momment. Anyways, thanks for these cool stats kd24.

In fact, this apparently popular set was what put peje at a disadvantage, killing its spinning starmie before it could kill skarmory and prevent hazard setup.

 

[Ojama]

Skarmory is now pretty common because of its ability to set up both Stealth Rock and Spikes very easily (the lack of Taunt users helps it a lot) and to take hits very well thanks to its great Defense and Sturdy. It is very interesting to see that people immediately wanted a replacement to Deoxys-D which actually shows that people don't want to use Stalls in this Metagame and want to be as much offensive as possible. It also shows that Weatherless Teams are being extremely popular which is pretty nice because these Teams are really funny to use and to face. I've played amazing games on yesterday's tour and those were always against Weatherless/Sand Offense (I'm not really considering Sand as a Weather or at least not like Sun and Rain).

Seeing both Jirachi ad Landorus-T at the top of usage stats isn't really surprising, at least for me, because they're just too useful in the current Metagame. I've never seen something as useful as Landorus-T in Offense, it can basically do whatever you need: Stealth Rocker, Late Game Sweeper, Pivot, Scarfer, etc. Jirachi is also extremely useful because of its resistance to Draco Meteor and its amazing movepool. It is without any doubt one of the 3 best Scafers in OU so this is really not a surprise to see it #1.

Since Melee Mewtwo is talking about Stall in his post, let's talk a bit about it. I used a Rain Stall during Sunday's tour and won but my opponent was using a really strange Team and so I won easily but it was in R3 iirc but be sure that I didn't use it for the rest of the Tour. While Stall was seen as a "safe playstyle" in ADV/DPP or even in BW1, it is absolutely not the case in BW2 and Peje's game just proves it. Masterclass brought the new Standard Weatherless Offense and absolutely wrecked him without any problem. If I had used my Rain Stall for the rest of tour I wouldn't have been in final for sure. This is just too risky to bring Stall in the current Metagame and I'm not sure if it's going to change...

Also if you guys are wondering what I was using, here are the Teams I used:

- Dragonite / Landorus-I / Keldeo / Latios / Jirachi / Scizor

- Terrakion / Landorus-T / Keldeo / Latios / Jirachi / Scizor

- Hippowdon / Keldeo / Landorus-T / Celebi / Jirachi / Terrakion (gr8astard's Team).

I'm now helping kd24 about this project so please pm one of us during the current Tour to give your Team(s) (only if you're in Top 16).

 

[dice]

Unlike Chou, I do believe that these stats do have merit. Despite the fact that tournament teams often use a gimmicky set or two to surprise the opponent, the teams are built to consistently win and you cannot achieve this without having solid offensive and defensive synergy. BW is often matchup related, but good play and a solid team can often overcome any kind of team you come across. The point is that these tournament teams are often built by better players than your standard ladder player on PS, so I believe they hold more merit in what Pokemon are better than others. For example, take the team made by my friend gr8astard that Ojama used; the team utilizes a Substitute / Hidden Power Ice / Thunder / Iron Head Jirachi to lure in the checks in counters for Double Dance Landorus-T and Band Terrakion. This is an uncommon, "gimmicky" set yet it works well in tournament play. I was recently on PS to mess around with the new sprites (the Cloyster is the best js) and the team I just fought on the ladder was Murkrow / Politoed / Ferrothorn / Toxicroak / Chansey / Kingdra (non Swift Swim). These kind of teams skew the stats, and which is why tournament stats can be more useful.

If anyone cares, on Friday when I made the Top 16 I used Ninetales / Dugtrio / Weavile / Xatu / Venusaur / Darmanitan in the battle.

 

[Darknut07]

Unlike Chou, I do believe that these stats do have merit. Despite the fact that tournament teams often use a gimmicky set or two to surprise the opponent, the teams are built to consistently win and you cannot achieve this without having solid offensive and defensive synergy.

I believe the crux of his argument was about statistical relevance. With such small samples, the error on each calculation is huge. To put it in a simpler way, Jirachi's usage is 45+x%, and in this case with a sample of 32 teams, x is quite large, which means we know very little about what the metagame is actually about. Referencing my earlier posts, Hippowdon (at first) having higher usage than Tyranitar had very little to do with metagame trends but was rather just an anomaly caused by the small sample (and 1 team that stayed the same with Hippo in it).

That said, this information can be useful if we look at it as how top players build teams that try and counter the metagame instead of looking for metagame trends like we've been doing. I don't believe the metagame is shifting towards Jirachi, but these stats can show us how good players are trying to use Jirachi's versatility.

Edit: Btw, before people say "as more tournaments happen, the statistical power will increase", as he mentioned, by this time the metagame will have changed and the older stats will now be irrelevant, and the problem of decreased statistical power for the new data will remain.

 

[Chou Toshio]

Basically, you have to ask yourself what these stats will be used for.

predicting ladder teams of good players:. Obviously flawed since the data isn't taken from the ladder and good players build differently for ladder and tournaments anyway.

Predicting Tournament Usage: Flawed because tourney team/trends have high churning, and the data is so limited, and so weak in explanatory power.

Tiering: You already know why this would not work

I guess the data could be use as Darknut suggested as a supplement to team building study, but we have better resources for learning team building, and there's only so much you can glean without seeing full team lists and sets.


So-- what for?

 

[Lavos]

predicting ladder teams of good players: Obviously flawed since the data isn't taken from the ladder and good players build differently for ladder and tournaments anyway.

i'd like to point out here that this claim is false. i'd like to think of myself as a "good player" and yet i always build new teams for official tournaments and then recycle them for suspect laddering. your blanket statements don't apply to everyone, so don't base arguments off them.

the way i see it, if kd24 and his crew are willing to go out of their way to compile some interesting tournament data for us, why the hell are we complaining about it?

 

[Chou Toshio]

^The point is that the population you're collecting data from and the one you're applying it to are different-- forget about the sample not being randomized in any fashion, it's not even being taken from the target population if you apply it to the ladder. Besides, we HAVE all the ladder data. Success against a small pool of opponents in a single tournament (which has a completely different format from the ladder) as proof of likely success on the ladder would be a very poor conclusion for instance.

The problem is not having more data-- it's people using that data incorrectly.

People drawing conclusions that are completely unreliable because of the incredibly low explanatory power of the data.

The problem is people getting a false impression of how the data can be used, and that fallacy being spread. If everyone were an intelligent "consumer" of data than there would be no problems.

Seeing respected/badged users/battlers in this very thread extrapolating far-fetched conclusions about the value of Landorus-T simply because it was #2 in this particular tournament is proof enough to me that users can and will be mislead by this type of data.

 

[MMII]

I thought the proposition was that this data be applied to the metagame as a whole? Maybe I'm misunderstanding you but you make it sound like the ultimate goal is to be successful on the ladder.

(I'm not sure what you mean about the Lando-T bit. It's an excellent Pokemon and the posts don't seem to claim that it is much more than that.)