Today, I have a bunch of awesome stuff for you. I am going to start with some amazing data, then throw in a tournament recap and a deck list because I know people love deck lists. I had written a long, boring description of my family’s trip to Massachusetts Regionals in May, but then I decided that no one really cared in the face of the Lysandre’s Trump Card ban. Suffice it to say, pre-Shaymin, Eggs was a pretty great play.
I got a lot of positive feedback on my last article — “No Kidding” — and I want to let everyone know that I appreciate the good feedback. Hopefully, you will find this interesting as well.
No point keeping you in suspense. It’s Wailord.
No, it actually is Wailord.
I could do this all day.
Actually, it is a bit more complicated than that. I have complained before that people do a lot of net decking when a deck performs well … all without having any reliable way of determining whether the deck that won the tournament did well because it was a popular play or because it actually was the best deck. If everyone is running Wailord at a tournament, the odds of Wailord winning are very high, regardless of whether it is BDIF. Also, the odds that Wailord performs the worst are equally high. Everyone is running Wailord.
The result is that to truly understand if a deck is performing well, you need to have the data on the total representation of a deck for a tournament. Of course, gathering the data on what decks everyone ran at a meaningful tournament seems somewhere between difficult and impossible. So I figured it would never happen. But then someone posted on Virbank a list of the decks being played by every single Master in Day 2 of US Nationals and I thought, “Wait a minute!”
So here was my methodology: I took every player in Day 2, which was kind of its own tournament, and graphed the number of match points they scored on Day 2 and the deck they played. Then I put in their Championship Points for the Top 100 US players (which takes you down to around 397). If the player was outside of the Top 100, I just put in 350 to simplify my life.
Then I plotted for each deck the number of match points the average player playing it generated and the averaged CPs for that player.
In this way, you can theoretically look at how each archetype performed while attempting to control for the influence of factors we don’t care about like “player skill.” Ha. As if that has any effect on outcomes! Unfortunately, this doesn’t control for poor deck construction, but what can you do. The good news is that all these people made Day 2, so we can hopefully say that we kind of controlled for that.
Before I show you the money shot, allow me to present some of the data. Here is each deck — and how many — from Day 2:
Toad/Bats – 10
Manectric – 9
Toad/Garb – 9
Metal – 6
Ray/Metal – 5
Lando/Bats – 3
Night March – 3
Wailord – 3
Groudon – 2
Kyogre – 2
Raichu – 2
Flareon – 1
Hippowdon – 1
Blastoise – 1
Lando/Garb – 1
Rayquaza – 1
Donphan – 1
Yveltal – 1
Trevenant – 1
Lucario – 1
Bunnelby – 1
Maybe you surmise from this that the “meta” was Toad, Manectric, and Metal with a long tail of “rogue” decks. That is probably about right. You can see here how, to make the data more statistically interesting, I didn’t lump Bats in with other things, but I did lump things like Manectric/Garb with other Manectric decks. I also broke out Ray/Metal vs. straight Ray because I thought there would be more straight Rayquaza when I started. You could break it down further, but I think the point is you need big, lumpy archetypes to draw interesting conclusions. When the n is 1 (each sample is one data point), then the only conclusion one could draw is that the best strategy for Worlds is to be Jason Klaczynski.
Here are the decks sorted by the average Championship Points of the people who played them (using 350 for those outside Top 100 US):
Metal – 596
Hippo – 581
Toad/Bats – 557
Blastoise – 547
Raichu – 536
Toad/Garb – 521
Manectric – 514
Groudon – 495
Lando/Bats – 493
Ray/Metal – 452
Kyogre – 442
Bunnelby – 440
Night March – 390
Flareon – 350
Wailord – 350
Lando/Garb – 350
Rayquaza – 350
Donphan – 350
Yveltal – 350
Trevenant – 350
Lucario – 350
So the best players, generally speaking, played Metal, Hippowdon, and Toad/Bats — potentially disproportionately benefitting those decks when you look at the results.
When you plot all the decks using my model, you get this cool scatter plot:
What conclusions can we draw from here?
- The deck that did the best with the “worst” players playing it was Wailord. No Top 100 player played it, yet it outperformed all but a handful of decks.
- Hippowdon and Blastoise performed the best on an absolute basis, but they were only played by one player each and those players were extremely skilled. It could easily be that those players would perform well with any deck.
- Raichu, Groudon, and Kyogre were sleepers. Those decks all worked well for people playing them.
- There is an unreadable pile of decks at the bottom. These are (from right to left):
- Wailord was the only deck in that bucket that had an n greater than 1.
I am inclined to ignore results from deck archtypes played by a single player. Similarly, obviously all of these decks and players were skilled enough to get to Day 2, so none of the decks are “bad,” but we see some obvious winners:
- On a relative basis, Kyogre was the only deck played by more than one player that outperformed Wailord. Now, “better” players played Kyogre, but not that much better.
- Wailord, Raichu, and Groudon were the best non-Kyogre decks played by more than one player, followed by Metal.
I did not bother to regress Metal and see how it would have looked if Dylan Bryan had not played it — he contributed a ton of match points and Championship Points. But maybe that demonstrates that good players play decks well. Similarly, I suspect that Toad/Garb would have been a disaster without Jason piloting his deck to the 8th seed.
Problems with the data?
- I think the Championship Points I used were post-Nationals and had been updated. This means that the outcome influenced the view of who we thought was skillful and not skillful, creating a correlation. Sorry about that.
- Some of this data is probably wrong — I may have copied down Championship Points wrong. Also, as I said, I made ad hoc decisions about lumping deck genres. And the random 350 Point thing.
Takebacks are allowed in Juniors, right?
True story: Nationals 2014 was the first tournament my son ever attended where he played best of three. After his first round, I asked him how he did:
Son: “I tied.”
Me: “Oh great, what happened?”
Son: “I won the first game, then we tied the second game.”
Me: “Uh … wait a second.”
Turns out both Juniors were confused and because they signed the match slip a tie, the TOs were not willing to unwind it. That sucks, right?!?
The nice thing about Juniors is that you assume they will never make these mistakes again and they have plenty of time. And that is pretty true, more or less. So we can laugh because they are getting better. Whereas if a Master makes these nervous errors, it is more tragic.
On to this year’s Championship. This was probably the toughest tournament ever (uh, in the last 2 years?) for Juniors to prepare for. As discussed in my prior articles, Juniors are much more reliant on studying results from prior tournaments than Masters or Seniors because they have a bed time and are forced to be well-rounded people by these terrible grown-ups that seemingly control their lives. The result is they have almost no time for testing.
With this in mind, how does a good Junior figure out what to play? We asked friends. Also, we hired a consultant. I had the benefit of a double stipend — despite only attending three Regionals this year, I had two children in the Top 100. Yay, Pokémon! So we took some of those funds and used it to hire a SixPrizes coach. We figured we would see how it was.
I am happy to report that the SixPrizes coaching program is pretty good for Juniors. We got some sound advice on gameplay technique and some good deck consulting. Unfortunately, I will say that every Master had much the same problem we had: The meta was still quite vague up until Canadian Nationals and even on Thursday, many Masters were still changing their deck. Getting good intel on the meta is a fight for everyone — even your consultant.
We were dialed in on Toad/Bats very early. My youngest had played Toad/Bats at Massachusetts Regionals and just missed top cut. When we saw Lando/Bats performing well, we were even more inclined to bring the Toad out to play.
The other card we liked in the post-LTC format was Aegislash-EX. People would be thinking about their Energy lines, but so many decks (Rayquaza, Raichu, Landorus, Toad) relied on Special Energy, which made Aegislash seem like a notable card. Also, we were drawn to decks like Metal and Yveltal where Energy recovery from the discard was strong. It seemed like Juniors that were slow to embrace the changing meta might come bringing a wave of Toad and Energy suppression and post-LTC, if you could ride that out you got an easy win. And of course, our first practice game on Thursday in the open room was against Piper Lepine (a notable junior) and she was playing Toad/Crawdaunt.
My oldest decided to play Metal at the last minute. Here is the deck list we played:
Pokémon – 17
Trainers – 30
Energy – 12
Really the only funny thing about our list was 2 Shaymin — in practice, we Ultra Balled for Shaymin first turn nearly every game. I am on the record validating the tremendous value of turning Ultra Balls into draw power. If Shaymin was prized, it would have been crippling for the setup of the deck. So we cut a Toad for a Shaymin. It made the deck super consistent and probably didn’t change our results.
Here is his tournament report:
R2: Vash Gotham – Virizion/Genesect – WW
This was revenge, as Vash had just beaten his brother in Round 1 (playing Toad/Bats), so that was nice.
R3: Kayden Shivers – Ray/Metal – WLW
R4: – Jon-Luke Modory – Lando/Bats – LL
Jon-Luke came in second at Nats last year. First game he had a Turn 1 Landorus with a Strong Energy and a Muscle Band and 4 Zubats on the Bench. So that was bad.
R5: Brayden Morin – Lando/Bats – WW
R6: Max Overdevest – Groudon – WLT
In the second game, Max had a Turn 2 Gaia Volcano with Mega Turbos — all basic Energy so he couldn’t knock it off or be saved by Cobalion or Aegislash.
R7: Road Godfrey-Robbins – Toad/Bats – LWT
He decked Liam in the first game.
R8: Henry Jung – Toad/Crawdaunt – WW
Liam played Metal instead of Toad because we thought there might be more decks like this. Easy victory for a deck that gets Energy out of the discard.
So his final record was 5-1-2, which put him at 24th place. He was happy as he never was below Table 13 for the entire tournament, but man, it is hard when you have 200 people and you cut straight to Top 8. If one of our ties had been a win, we would have been right there.
One of my new goals for Worlds is to try to prepare my kids to have a mature conversation with their opponent about preventing the third game from ending in a tie. Both games, Liam was ahead on Prizes at time and needed an extra turn or two to complete the game.
Further, I know we didn’t even get the worst of it. He came in 24th and was 2 match points out. That means EVERYONE basically bubbled out. And it’s easy to say, “Hey, they are Juniors,” but most of the people ahead of my son will be Seniors next year. Many of them viewed this as their big chance. Regardless, 96% of the kids that had flown from all over the country were done with their competition by early Friday night. It’s a kid’s game! More Pokémon, please!
One last complaint about Nationals (and I made all these complaints to the staff): SELL SOME FREAKIN’ SHIRTS. If you don’t want to give them away, no biggie, but my kid wants to wear his Nationals shirt. Hook us up!
If I had known that they would have no shirts, I would have printed my own and made a killing. Just put a Rayquaza on the front and start cashing checks.
I look forward to seeing everyone at Worlds. Please come introduce yourself! I will be the Pokédad trying to get the kids to eat something.