In Defense of Net Decking

Thanks for the many kind words regarding my prior article. It has been a pleasure to bump into so many people and talk a little math since my original article came out. I look forward to seeing people at Virginia Regionals. I hope you have all been doing your math and building better decks as a result.

It has been a fun City Championships for my little guys, although we were too busy enjoying Santa Claus’ visit to hit the marathon circuit. But it did give me time to reflect on how helping my son excel at Pokémon has evolved. Today I want to write a bit about a non-mathematical concept that has also improved my deck construction habits: net decking! I come here to praise net decking, not to bury it.

Preparing for Tournaments

pokemonscreenshots.tumblr.com

As a Pokédad, I am frequently explaining to new Poképarents the concepts of Pokémon. I generally distill success at Pokémon down to a three-step process:

• Good cards: Given two people playing, best cards win.
• Skill: Given two equally good decks, most-skilled player wins.
• Luck: Given two equally good players playing equally good decks, it is still a card game and luck plays a role!

Good cards at the highest level, where it can be assumed that every player has access, more or less, to every card, is equivalent to calling the meta. Having said that, I would contend that this is at most 50% of the importance of the game. There are a couple of things you want to work on in preparing for tournaments that we can call the skill-based portion of the game:

• Identifying the best meta deck to play
• Identifying the best card-by-card configuration of said deck
• Improving in-game decision-making and technique

Each of these things can be cited as the source of a few wins. Interestingly, despite the overwhelming focus on decklists, there can be little doubt that the last of these, in-game decision-making and technique, is far and away the most important of these skills. Almost no decklist decision guarantees victory, yet tournament after tournament, the players who have consistently performed well seem to continue to generate outstanding results.

On the one hand, this could be attributed to their superior deck-building skills and understanding of the meta, yet we see incredibly widespread variance in decks but far less variance in the results yielded by these decks from a small group of players. For example, if you look at Ft. Wayne Regionals top 8 in Standard:

• Dustin Zimmerman (Virizion/Genesect)
• Kyle Sucevich (Aromatisse/Mega Kangaskhan)
• Joe Baka (Yveltal/Seismitoad)
• Alex Hill (Donphan)
• Nikolas Campbell (Donphan)
• Frank Diaz (Donphan)
• Jason Klaczynski (Donphan)
• Henry Prior (Donphan)

While you can see the success of a strong meta deck (Donphan), you can also see that there are 4 different decks in the top 8. Also, there are names that everyone in Pokémon knows well appearing here, some playing Donphan, some playing other decks. There are many players here that simply produce consistently good results regardless of the deck choice. The prior weekend, Jason Klaczynski had been in top 3 playing a completely different deck.

Brit Pybas now maintains a list of his perceived ranked value of players. I don’t think he is making a list of great deck builders. If you handed everyone the same theme deck, these are the people that would march out of the tournament on top.

Anyone that has played the game and flubbed a simple order of operations, made a terrible misplay, or seen someone make an incredibly tech play for the win knows that in-game skill is a huge difference maker in the game. Similarly, as I am acutely aware, most Masters consistently malign Junior and Senior tiers, but I think with the prevalence of net decking, none of these Masters think that it is because they are playing with worse decks, they think that it is because they are not skillful in-game players. There is probably truth to this. There is a reason that Juniors games take about half the time of a Masters game and I suspect it isn’t because Juniors are smarter and thus quicker. (Although my Junior would happily tell you that he is much smarter than me.)

As a sidebar, one of the things that aggravates me about post-tournament reports (and I recognize this is challenging to rectify) are that they report the top 8 decks like they represent the best plays in the meta at that moment. More interesting would be to understand what proportion of the total decks played in a given genre represents so you could compare top 8 outcomes to the hoi polloi.

For example, if I told you that I had studied the results from the rest of the pool at Ft. Wayne Regionals and it turns out that every other player outside the top 8 at the tournament was playing Donphan also, that would indicate that Donphan, rather than being a successful meta choice, was popular in the top 8 due to the sheer number of players playing it rather than some spectacular BDIF or sweet meta call. Virbank City and SixPrizes both regularly see decklists from average performers at tournaments that are a card or three off of decks that performed very well at the tournament. I suspect we would find, if we could unwind this, that the difference in outcomes is probably more closely related to in-game skill than meta calls or individual card differences.

Deck Construction Practices

wchfh.org

As the recent success of Donphan has demonstrated and the tremendous response to Dylan Bryan’s dissection of Donphan strategy here on SixPrizes demonstrate, building the right deck can make a huge difference. So picking the proper meta deck is incredibly important. Of course, every player would say that deck tuning and making sure that your deck has the right cards for the strategy and your respective playstyle are just as, if not more, important. Dylan’s article demonstrated that it was not just “DONPHAN” that made that deck work, but a specific set of cards used in a specific set of situations that he anticipated.

As a Pokédad, over the last 8 months, I have asked literally dozens of players how they playtest and my specific question was, “How do you test to compare how a one-card change affects deck performance?” If you change one card in a deck to improve a matchup against another specific deck (for example, the mirror), it is easy to test how much that card improves the matchup against the deck it is designed to counter, but teching in a card typically comes at the cost of decreasing general consistency. Testing how much consistency has been lost seems like it would take many, many games to detect given the minor overall effect of the change in the broad scheme of things. Universally, the answer was, for better or for worse: play lots of Pokémon.

So testing decks involves a two-step process of determining the general meta concept you want to embrace and then optimizing the individual cards of the deck, requiring meta play and then iterative testing through an optimization cycle.

Unfortunately, everyone has time constraints. Myself and many Masters are constrained by personal and professional responsibilities. My kids’ playtesting time is constrained by their scholarly obligations. Many Masters, for that matter, have similar, if even more intense scholarly responsibilities. Or maybe you are playing poker full-time. All these things constrain time for testing and optimizing decks.

So how do you identify a good deck to play in the meta? I would posit that you have to try a bunch of decks. Every serious player in the game has probably played a couple of games each with most of these decks:

• Metal
• Fighting
• Donphan
• Yveltal
• Virizion/Genesect
• Plasma
• Pyroar
• Seismitoad
• Gengar
• Night March

Conventional wisdom is that this helps you understand the deck and it helps you understand how a deck you might play would play against that deck. Hence the importance of testing a wide variety of decks.

Net Decking

Bulbapedia

I read SixPrizes comments. I read Virbank City. I read other sites too. When people ask for decklists, the comments turn pretty aggressive. Let’s be frank: “d— riding,” “sh—-, stupid player,” “too dumb to build their own list.” I think these are some of the nicer responses you get.

This begs the question: should people build their own list? I would posit that the answer is no. I net deck all the time. I LOVE NET DECKING.

With the release of Phantom Forces, we needed to test our deck against some of the new decks in the format. Now, we could build our own Metal deck, but then we would have to spend time discovering the flaws in the deck before we would have a deck strong enough to credibly test our existing deck at a competitive level. Honestly, who has the time? I don’t want to fine-tune and optimize a deck that I am not interested in actively playing. I want to playtest this matchup. I just need a representative deck that other people have already thought through. Go to SixPrizes, net deck, voila.

Is this bad? Does it imply that I am short-changing my own Pokémon development skills? Possibly. Maybe I would learn something about the broader nature of Pokémon training by spending a few hours on a Metal deck that I will never play at a tournament, but I would rather build a Metal deck and work on in-game skill and optimizing the deck I want to play. Given time constraints that does not seem unreasonable.

Trade-offs happen and this is why net decking is good. Should I playtest against a Night March list that I threw together tonight and is terrible, or should I test against one that Kyle Sabelhaus has spent some time fine-tuning? I think I will learn more about how my deck performs against Night March using Kyle’s list than mine. Using net decks, I can test against three or four meta concepts in a few hours, where I would only test against one or two, and test less effectively, if I was not leaning on the Interwebs.

The power of the Internet is that I don’t have to figure out why rainbows have different colors. Google it. Would I be better if I figured it out for myself? Sure. But I don’t have the time.

The other side of the coin is in testing new deck concepts. If you think the meta is such that a particular play is the right one — “Shiftry is the BDIF!” — you will need to test your new idea against a variety of other meta decks. Should you be spending iterative optimization cycles on preparing decks to test against or optimizing Shiftry? In a time-constrained world, there are only so many decks that we can optimize at a time.

Net decking is an accelerator to testing and in a world growing rapidly more time sensitive, who doesn’t need that?

Iterative Optimization

pokemontime.tumblr.com

Many people would say that their problem with net decking is not really with net decking per se, but with people who play decks straight off the net without changing a card. Do these people deserve to be derided or do they deserve our pity?

Without iterative optimization and testing, they are relying almost entirely on in-game skill and calling the meta to see their way to success in a tournament. Making the small tweaks to a deck that optimize it to support your gameplay style or local meta are the differences between success and loss in some rounds.

Fine-tuning a deck comes from playtesting and taking time to consider the deck. Comparing your deck construction to other parties’ deck construction can be an important input into thinking deeply about how a specific strategic concept works. This is another area in which net decking can serve a serious deck constructor well.

In a recent SixPrizes article, Ryan Sabelhaus walked through his deck and how it varied from several other decks in SixPrizes. I don’t think anyone would say that Ryan is not a serious Pokémon player at the absolute top of the world’s most skilled Trainers. Yet here he is dissecting net-decked lists and discussing what he did differently, how he tested them, and what worked and didn’t work.

Similarly, despite Brit’s impassioned argument for how net decking is lowering the collective Poké-IQ of the world, I think you see mentions all the time on Facebook and SixPrizes of how people got lists from Brit that they used card-for-card at a tournament where they saw success.

That is OK, but let me be clear: I believe net decking or getting a list from a friend is fantastic for building a deck to test against or testing a potential play against the meta, yet very, very mediocre for building a list you intend to play. Only through iterative optimization can the perfect list be arrived at to support your in-game play personality and keep up with the meta.

Tournament Report? What?

I want to walk through one example of how I optimized a deck for a tournament I played at recently. Being a Pokédad, I rarely play, but I had a sick kid so I only had one kid with me at the tournament, meaning we had enough N’s for me to play also. Here was how our thinking at the Halliburton house had evolved:

We had wrapped up Furious Fists playing a list based on Brit Pybas’ Donphan/Raichu list, but had taken out the Raichu as the meta had swung to all-Donphan all-the-time. With the resurgence of Yveltal, I wanted to bring back Raichu, while keeping the Donphan to counter the Plasma and Manectric decks I expected to see.

Unfortunately, the list we had originally used was pre-Phantom Forces and we wanted to add a thick line of Robo Subs and VS Seeker, along with a Lysandre’s Trump Card. Finally, with Phantom Forces introducing Aegislash-EX and bringing back Enhanced Hammers, we would like to add more Energy. But this list is pretty tight! What to do?

I never imagined I would write an article that had a decklist. I am no Brit Pybas or Dylan Bryan! But here you go:

I felt like this was the skeletal remains of Donphan/Raichu. Deep thought about the prior list and iterative testing led me to make a few changes from Brit’s list:

1. No Mewtwo: Deck just didn’t have space. Would have been nice. Many people also run Dedenne FFI to address similar problems, but I figured Raichu would have to do.
2. Kept the Suicune: I recognize few Donphan decks run Suicune, but when you have that many DCEs and a big Korrina line, it isn’t hard to make sure you can retreat and that extra 10 HP can be a life saver vs. Baby Yveltal with a Hard Charm. Most Donphan decks today don’t run 4 DCEs, but when you do, a 2-Energy Retreat wall is a luxury that can be afforded.
3. No Reshiram: This was simply consistent with the evolution of the meta.
4. More Korrina: I didn’t have room for a lot of Ultra Balls and (as my prior article attests), I didn’t want to have draw problems early, so this would ensure I got my Donphans out consistently even though I only ran a 3-3 line.
5. Less Colress: Only had room for one, but I had VS Seeker to bring it back repeatedly, which I did.
6. Less Lysandre: I had the choice of 2 VS Seeker and 3 Lysandre, which would mean I could draw into a Lysandre sooner, or more VS Seeker. I like more Seekers generally to improve draw probabilities. Further, I always feel like if you run a Lysandre’s Trump Card and you don’t run at least 3 VS Seeker, you end up having to try to conserve VS Seekers. This is not dissimilar from running Jirachi-EX and only running 1 Ultra Ball. You are taking up a lot of deck space without dramatically increasing your deck flexibility. There is a 50% chance you will see your Lysander’s Trump Card and one of your VS Seekers in the first 30 cards you look at, but you won’t want to play it then, probably. If you feel obligated to conserve your last VS Seeker for virtually the entire game, then you are not getting much benefit from those two cards. Similarly, the more Ultra Balls you run, the more benefit you receive from running Jirachi-EX.
7. Added Trump Card, Robo Substitutes, and VS Seekers. Mission accomplished.
8. A balance of Bangles and Muscle Bands. Without the Mewtwo, there was little reason for me to overweight toward Bands and Bangles help me hit better math on Manectric KOs. (1 Strong + Fighting Stadium + Bangle = 1HKO on a Mega-Manectric)
9. Less Float Stone: One of the great things about this deck is that it has so many free retreaters: Raichu and Hawlucha. Suicune and Kyurem were the only things that needed a Float Stone.
10. Less Ultra Ball: We had playtested Donphan for weeks with 1 Ultra Ball. Would have been nice to have more ways to build Raichu, but I thought of him as secondary to Donphan anyway and with the 3-3 line, I was likely to draw into the cards regardless. This worked out fine.
11. Escape Rope instead of Switch: In playtesting, Rope turned out to be more useful than just a plain Switch as it could get you out of a Status Condition or make the opponent make tough decisions. With few lasers in our meta, Escape Rope was probably more useful.
12. Similar Energy line but didn’t worry about the Water. Would have liked to add one more Energy, but it didn’t happen.

With these relatively thick lines, I saw consistent draw cards, Pokémon, Energy, Tools, and Robo Substitutes. The only time I was unhappy looking at my cards was when I drew completely dead for 4 turns at the start of a mirror match vs. Dylan Dreyer.

Problems with this deck: As Brit saw, it frequently loses the mirror because Raichu is Weak to Fighting and it is hard to power up a Wreck consistently with so little F Energy. I went 1-1 vs. the mirror. Conversely, the deck did its job against Speed Lugia and a Manectric/Black Kyurem deck.

The result: Pokédad is top seed going into top 8 at Frederick Cities:

I felt lucky and fortunate to have my name on a list that includes some of the people that I think of as among the best players in the game. Many of these people have provided thoughtful insight and helpful coaching to both my sons and me and I was grateful to have the opportunity to play some fun games with them. I am scarcely worthy to carry their Poké Ball!

Of course, it is worth pointing out that second through fifth all ID’d their way into the top 8 and many would have had better records than I if they had played it out. I was the 3-1-0 that got down-paired so I had to play my way in. I made my way to top 4 before losing to Jimmy Pendarvis playing Yveltal/Hard Charm/Shadow Circle, who went on to win the tournament.

To be clear, that was a round where I felt like I had a reasonably strong deck for the Yveltal matchup, but I simply had inferior in-game skill. There is always room for improvement and he simply played his deck so much better than I played mine that victory proved impossible. Frankly, the prior round against Dean Nezam was another round where he had far more in-game skill, but his decision to run a single Kyurem PLF with his Mega-Manectrics proved costly.

This demonstrated my fundamental message. Success relies on predicting the meta, making card-by-card deck optimizations and in-game skill are the three keys to success, with each skill being more important than the prior. (Dean told me that he had taken out the second Kyurem from the original decklist he had been given because he figured he lost to Donphan either way, but combined with his superior play skill, that would almost certainly have been enough to swing the matchup in our games. Another example of great players taking lists and thinking constructively about them, although in this case it served him poorly by assuming he would play someone with skill and talent rather than myself.) Those who choose simply to net deck neglect the second habit of success in Pokémon and are unlikely to see the success of people that work hard to evolve their decks.

I think the minutiae of better in-game decision-making is one of the things that is a huge opportunity for future articles and the challenge that every Pokémon player works on every time they play. I look forward to reading more about it. For now, I hope this article has helped you think better about how you can take decks you find online and go through the process of iterative optimization to make those decks your own and generate the best possible outcomes.