Failing at Worlds

Today's Soundtrack: https://open.spotify.com/track/3S9ztrnCBKQ8pIL9iLQdSM?si=c2c321579fe34f10

"I'm not trying to panic - I'm trying to make an actual evaluation of how bad this is, because I have to play in Worlds tomorrow, and it's useful to know how difficult this matchup is for decisions like whether I need to make them have an answer they're somewhat likely to have, whether I want to be trying to go deep into the game with them or lose hard and early but have a chance."

I'm writing this defeated. It's not a good feeling, even if it's a necessary one. While Sanctum overall had a better performance this year at this Worlds than last year, results aren't always the most accurate reflection of process, and process is what you need to focus on if you want results in the future.

Another team brought a broken deck to worlds, and we failed to find it. Even beyond that, we brought decks that were optimized for a far different metagame than the one that actually showed up, even ignoring Lessons.

If you'd asked me the day before the metagame breakdown came out what decks I was most preparing to beat, I'd have told you Izzet Looting (4th most popular), Simic Aggro (7th), Dimir Midrange (not even in the breakdown), Jeskai Control (5th), and maybe Bant Airbending (3rd). Sure, a couple members of other teams might bring Otters, I'd have admitted, and yeah there definitely could be a broken version of Izzet we hadn't discovered. But fundamentally we missed the top 2 decks. Small fields with lots of testing teams can be super swingy like this, but still, this was a gigantic miss by any measure:

Today I want to explore what went wrong with our testing process for this tournament (and what went right too -- even though overall I'd say we came up short of what was needed for success, there were absolutely some wins along the way).

I'll be writing a mini-guide to the new iteration of Otters that we brought and commenting on its positioning as well as the list others brought with 4x Boomerang Basics, but first I wanted to take this moment where I'm reflecting on our process as an opportunity for a deeper dive behind the curtain of a testing team -- Let's unpack the work that went into us registering such a diverse spread of decks for Worlds, and where we could have done a better job of finding the actual best deck and preparing to beat the decks people actually brought to the tournament.

Before the Ban

This piece is going to feature more excerpts from Sanctum's "Testing Notes" discord channel than the rest of my post-tournament deep dives. Partially this is because I'm trying to reconstruct the path that led us here as I write, and partially because this channel was far more active and fleshed-out for this tournament than for a normal PT. I absolutely loved practicing with this Worlds team, and part of what I want to show today is that is certainly wasn't a lack of effort that led us to the wrong conclusions about the format, but a few subtle misplacements of our efforts.

One disclaimer I'll give is that despite the amount of stories I tell here, I'm necessarily going to be leaving out the VAST majority of stuff we explored and talked about throughout the month and a half before Worlds. There were 30,000 messages sent in the Worlds server from October 1 through Worlds -- to give a sense of scale, 30,000 words* *is roughly the length of a hundred-page novel. So however many words are in the average discord message, that's how many novels I'm trying to summarize here.

When we started testing, we knew that there was going to be a very tight turnaround between the upcoming Vivi/Cauldron deck being banned, so we wanted to get in as much "pre-work" as possible. We started out by just playing a lot of internal standard without either Vivi or Soul Cauldron. One of the first decks we started working on was Otters -- not strictly because we were so familiar with it, but because the previous worst matchup, mono red, had had some cards banned, the current worst matchup was Vivi/Cauldron, a deck about to get banned, and the core of the shell was still intact and very powerful, while being largely absent from public results.

The early testing we did with otters was pretty mixed - it still struggled against red, but the matchup wasn't nearly as terrible as before, and the other matchups seemed reasonably good.

We also were testing a good amount with Dimir and Mono Red in general, as these seemed pretty likely to be the day-1 baseline pillars of a post-Vivi/Cauldron world:

We also tried the old standby of adding Innkeeper's Talentto the red deck to add some resiliency to removal and ability to grind, while Jesse and Casey had started to brew with the package of Icetill Explorer, Esper Origins, Overlord of the Balemurk, and mana dorks:

Meanwhile, Jason was tuning Otters to see if we could make it better into the red matchup. In classic Sanctum fashion, this involved adding another color to the deck, this time white for Seam Rip and Brightglass Gearhulk (which can tutor up both the Seam Rips and Stormchaser's Talents):

Besides this brewing though, a deck was emerging out of the rest of our testing that seemed much more resilient than we'd expected: UR Vivi-less Profts. Izzet cards were still pretty broken, it turned out:

... which resulted in a consensus forming:

But then this cute little guy got spoiled....

And boy did it feel like a powerful inclusion in Otters...

Soon the only two decks putting up clearly good results in our testing were Badgermole Otters and Profts:

The Icetill deck that Jesse was now soft-championing at first seemed like it had an unforgivably bad profts matchup, but then was able to solve it by shifting away from a white splash for Sheltered by Ghosts and Seam Rips for the red matchup, back to a pure Golgari base with some Season of Loss as a hybrid wrath/finisher.

We were still playing a lot of mono red as the aggro deck of choice, but it started to be supplanted by this Simic Aggro deck.

Paul was the main champion of this deck:

... and while it didn't seem that strong or tough to beat on paper, it proved very difficult to beat convincingly in testing. The one deck that did so somewhat convincingly was Otters, by taking advantage of Simic's lack of interaction and slightly-slower-than-mono-red clock:

So this is where things sat before the ban announcement on November 10.

We assumed one of Vivi and Cauldron would be banned, presumably Vivi, and were hoping very strongly that Profts would also get the axe, so that we could at least consider playing something else.

We had multiple other decks we were pretty optimistic about -- the GB Icetill shell, Simic Aggro, Otters, and possibly some version of red aggro as long as Simic didn't become too popular.

Post-Ban

Thank god, they banned Vivi and Profts!!!

.... and Screaming Nemesis???????

Well, THAT was pretty unexpected.

I mean, this didn't come completely out of the blue -- Nemesis had been a powerful card whose play patterns people complained about for a long time, but it was pretty surprising to ban a card out of a deck that hadn't been tier 1, let alone clearly good before. But we had been so locked in on hoping that they actually banned profts, this came as a surprise.

And basically every factor here was clearly in the direction of things being good for Otters...

• Profts, the 'tier 1' deck that was structurally well-positioned into Otters in our internal meta just got banned

• The other card that got banned was not only a flagship card in Otters' other worst matchup, it was the card Otters had the toughest time answering

• The decks that looked to be "level 0" going forward were badgermole cub-style elfball decks, which Otters had already shown to be quite favored against

(we started calling Otters 'Rats' as a half joke, because the deck now played 4 Song of Totentanz, and we were wondering if in a Badgermole world we might end up cutting the Thundertrap Trainers)

That being said, while the joke is that Sanctum always plays Otters and is too high on the deck, and while most of this piece so far has been talking about Otters, and while I did end up playing the deck at Worlds... my reaction to this news was less "this is awesome!" and more "oh man, am I actually going to have to play the same deck I played last year again, when everyone will be expecting that from us? I hope we can find something new and better".

After our post-ban constructed meeting, I took on the task of trying to find the best nemesis-less red deck going forward, while others on the team naturally wanted to try out Otters. This led to us re-checking the Otters vs. Red matchup, this time without nemesis...

However, this version of the red deck, basically just the same deck but without Screaming Nemesis, felt pretty horrible and I was pretty confident it wasn't the best red aggressive deck that could be built. We did some testing with the leyline red deck that had just won the spotlight, and while it felt swingy and inconsistent, it was definitely powerful, and seemed to crush Otters...

.... or did it?

After this set, there was a decently long discussion in the 'Rats' channel about how Otters wanted to play the matchup based on our learnings from last worlds, and how it was very important to save your interaction to hold up during and shut down their combo turns, even if it meant taking some damage early -- exactly the opposite of the way you needed to approach the traditional red matchup, where you wanted to win the early turns of development and worry about taking incidental damage, otherwise you'd get finished out of nowhere with a Screaming Nemesis or Nova Hellkite.

Cft seemed pretty insistent that we re-try the sets, and...

Meanwhile, some of the work Jason and Talia (Bael, a longtime Sanctum member and control theorist who helped us out for Worlds) on a control deck with Legend of Kuruk as its finisher had been doing was showing some success, despite the deck playing the incredibly stinky card Virtue of Loyalty:

I have a ton of respect for Talia especially when it comes to faer's ability to build white-based control decks that skew more board-presency than most. Talia was a key architect behind the Serragon deck that Jason went 8-2 with at PT DFT this year (notably, other Sanctum pilots struggled with the deck), and just generally has some of the best deep understanding of the game I've ever seen -- so despite some of the card choices being stinky on the surface, I was paying attention to this space, especially once Simic & other Ouroboroid decks started to see a lot of play online and the general idea of playing a deck with wraths in it became appealing...

A lot of work was being done at this stage on how to build the best Badgermole/Ouroboroid deck we could, as this pairing was starting to be the clear most powerful thing to do in the format. Somehow, this brought us back to the card Virtue of Loyalty in a bant flash shell:

We were very high on this style of deck!

We also found the same "GB Ouroboroids" deck that some folks ended up bringing to worlds in results from a big Japanese tournament, and Mora was doing some exploration to try and see if this type of 'take the best game action as many times as possible' deck was what we wanted to be doing:

Meanwhile, cft had been exploring some Izzet shells, in what we all agreed (as did the online meta) would definitely still be a space with lots of viable decks going forward. They hadn't found anything that impressed them (it's hard to impress cft), but they were still quite insistent that there was likely a tier 1 deck, if not a 'best deck' in this space.

I largely believed this, and my attempts to find it led me to a version of Prowess that was playing Talent-Thundertrap-Stock Up as its card engine (using Stock Up and Ral/Enduring Curiosity/Steaming Sauna instead of Riddlers and Splash Portals for card flow), and used Valley Floodcaller as its 'output', staying more low-to-the-ground and being able to play more cheap interaction (8 bounce spells between playsets of Boomerang Basics and Bounce off, plus Torches) to fight the Badgermole decks. I tried to prove this deck's worth in a marathon testing session against cft on a variety of other decks:

Meanwhile, cft remained high on the Splash Portal version of prowess, which I definitely thought was powerful, but felt too clunky to really be a 'best deck' to me.

Around this time, we had another meeting about Standard. I don't remember anything that was said here beyond various summaries of everything I've mentioned just now, but what did stick with me was cft's text-based thesis on the format. If you're wondering how close we got to Lessons and what our general theory of the format was about a week out from deck submission, I think just putting this here in its entirety is as good a summary as you're going to get:

(Everything that follows is from the great cftsoc. The numbering is by individual discord message, and I haven't changed any of the other formatting)

ok

Here is my General Theory of Standard. I tried to wake up before the meeting and write this out but it took too long oops

Mole

• i guess there is enough hype around it that the format starts with the mole. the primary property of the mole is that the mole structure is very fast if they don't do anything about it. that’s kind of it. if they do do something about it it’s much slower, but you’ve forced your opponent to use an interactive spell and then you can maybe cast your big spells naturally.

• a natural payoff for the mole is ouro. it can appear on t3 even thru one dork being removed, it generates immediate impact, it wins reasonably quickly, it’s resilient to some kinds of interaction, it can win when cast naturally without a huge amt of help, it works well with your card that generates two bodies, etc etc. so the main mole deck is simic, i would say i like the blue cards, mockingbird and drowner seem nice in mole mirrors, repulsive is very nice vs various things people can try to do to you, but i would say there’s definitely a lot of work that could be done for this stuff, a lot of possible refinements, a lot of possible interesting cards in other colors, i think splash lively dirge is pretty interesting, etc etc. there’s the golgari deck, the golgari deck should maybe just be spliced into sultai, the gruul deck is kinda interesting in that i felt the delirium and ouro stuff were pretty disconnected but it does like ask some different questions like eg pyroclasm is horrible vs all the delirium creatures, you would like bounce spells for the pump but bad vs most of the rest of the deck, etc. like you could also think about, like green cards plus idk red cards or something, not even delirium, cast nova hellkite and shock and whatever idk, some stuff to consider…

• then there’s also nonouro mole maybe. primarily this is appa mole rn. i mean this is also an interesting deck in that it asks some different questions, shocks clasm not very good vs appa etc. without many useful blinkables though you’re kind of weak to doom blade among other natural structure issues… idk. the mole structure just has fundamental problems. your deck is naturally lands dorks and payoffs, you need to have a correct mix of stuff to win, you have to mulligan, then your opponent can try to stop you at either end with the dorks or the payoffs, the best thing they can do is first stop your dorks and then your payoffs are generally quite slow and they can figure out what to do about that at their leisure. i think in general the mole is a scam. i suggest other people not to believe me though because there’s still work to be done in the space.

Blue Small

• the primary property of the blue small decks is that they will never run out of stuff. each of these decks can play like 8 of the broken blue card draw spells (stock up, winternight, riddler, kuruk, accumulate wisdom, etc etc etc) and supplemental selection (fomo, trainer, stormchaser’s, boomerang, etc). they will never run out of resources. this is a really broken property to have, to always have stuff to do and indeed with the immense amt of selection to always have the right stuff eventually as long as the right stuff is in your deck.

• the question becomes how to effectively convert these resources into, well, doing something. obviously the natural pairing is cheap interaction and that’s really convenient, because that’s a nice formula for beating mole - interact with early development, refuel, get on the board and interact with the payoffs when they make it there. but then there are questions like, how are you planning to attack for lethal? attacking for lethal is at least a little important to defeat a variety of things; prominently in the mirrors no one will run out of stuff so you have to have a plan for how your stuff is gonna kill them better than their stuff will kill you. there are a lot of different existing variants and i definitely don’t think i know enough about each of them yet but i’ll discuss them briefly:

• Izzet portal gets to have huge selection for the right spells with trainer blink, it’s the most consistent at defending itself if it has the right spells in its deck. however its ability to attack for lethal is the weakest and especially slow if your opponent can remove a riddler. probably the worst into beeg decks (i’ll discuss this later). i think i’m off this one for now despite it initially feeling good.

• izzet lessons plays the most powerful card draw spell accumulate. i mean all the card draw spells are broken but there are kind of degrees to this stuff, accumulate is really wild like not far from cruise at all, can play gran-gran to be really efficient in the endgames, combustion technique being true doom blade is a huge improvement, etc. one major weakness is that gy hate is actually effective vs you, which could be a dealbreaker, idk, i would say i’m still quite interested in exploring this more. similarly to portal though not totally clear how you’re supposed to try to attack for lethal.

• izzet discard claims that it will create board presence and attack for lethal by casting cheap creatures with large amounts of power and toughness. it’s a very good theory! maybe the best theory available from the ones here! this is maybe the stock deck that makes the most sense to me. i will say that the boomerang furnace stuff was a lot weaker here than in the other variants i felt, and there are definitely awkwardnesses revolving around being completely reliant on riddler for card advantage and not always being able to keep both interaction and creatures as you’re going places, and then maybe it’s weaker into people that are like “good vs creatures” like gumming up the board or wrathing you or something, but idk. i mean i just don’t know how good the creatures are in general at attacking people to death because it’s very nebulous, this deck needs far more investigation.

• dimir bounce claims nwtr and doom blade are better than red removal. it’s pretty believable. bouncing grim bauble is also pretty effective vs the cub decks i felt. could lean more into the bounce stuff (splash pixie) or away (cut fiso) idk i have no idea how good tbju is etc etc. i mean as the section title suggests all the important parts of this are blue, it’s really just the same concept etc. i will say losing red makes the normal dimir mu a bit scarier though definitely (it could be fine, idk). idk this deck felt a bit nice

• otters claims that it will attack for lethal by using the power of enduring vitality. it’s a really perfect fit, if you have all the resources you want a sticky powerful mana source, it works well with the cards you already want to put in your deck, and it ends the game completely and totally. the only flaw is that there are many more Problems. like the UR decks are just so clean, pristine mana, otters almost can’t play boomerang because it can’t have U into UU, has to play a ton of lay of the lands for its trainers, and then they never draw a useless song or floodcaller. but i mean the payoff is definitely there i think, but there’s a lot more stuff to explore still even, i mean if you manage to cut the trainer things really open up definitely and you maybe don’t have to play a ridiculous amt of taplands but @ jasoniltg definitely right that the trainer helps a ton with development towards combo whilst not losing tempo. idk it’s very complicated. lesson version? splitter version?? idk this is maybe still my frontrunner in my head but the Problems may be too Problematic. i mean also the mirrors where your opponent has a lot of ways to disrupt your combo cards might make it not nice, even though having any combo at all is nice and it's not like extremely easy to disrupt or anything, it seems quite complicated and needs to be tested

• and i mean there could be infinite different structures of this stuff. i mean there’s also the ryan UR floodcaller deck or whatever. but like again all the important cards at the core are blue, could explore various different color combinations (though the boomerang does kind of want 2c for U->UU), maybe explore various different effective cheap creature strategies or something, idk.

• i will say that one might think, well the mole structure generates mana, the blue structure generates cards, can we just combine them? i feel we tried this with some early builds of badger otters and simic and i don’t think it worked that well, it’s just the same problem as the aforementioned mole problems, if they disrupt your early mana then you have some card draw and it’s hard to catch up with dorks rather than interaction after you cast your card draw, i don’t think it works that well. i mean maybe you should think about other ways to generate large amts of mana. icetill (+virtue?) in this kind of structure maybe? put omni in play? idk

Tempo

• the third deck that people will play a lot maybe is dimir. from a deck structure perspective, dimir is bad against small creatures because its central cards kaito and curiosity are severely weakened. as such, it’s naturally bad against mole and blue small. you can gain enough winrate with a good sideboard plan such that you could maybe make it to 45% or something so i would guess a lot of people stay on this deck, but i would be really surprised if it’s a good choice.

• you can try to get better at its problems by like doing specific stuff like avatar’s wrath or something. it feels pretty unnatural to me though, like it’s fighting an uphill battle against the natural structure of your deck. the tempo structure is appealing when its good cards are naturally well-positioned in the format; i don’t really see a reason to try hard to register them at the moment.

Beeg

• most of the remaining vaguely serious decks are kind of in the Beeg category. because the Blue Small decks are not that good at ending the game, you could argue well i’m going to do something Beeg and get them before they manage to figure out how to attack for lethal, and then separately you could say things like, well interaction followed up by wrath is good vs the mole ouro decks, etc. so you get these decks like icetill, moleless appa, living end, ctrl kinda… i would say, like, playing the cheap cards is just inherently stronger than playing the beeg cards when you don’t run out of cards, and also beeg cards are like much weaker into countermagic than small cards, and then with fewer cheap cards you become naturally weak into dimir rather than naturally strong. this isn’t to say there isn’t any exploiting to be done, living end is probably good if no one ever casts gy hate, icetill is genuinely good vs the UR decks as currently built, appa does like do some stuff idk… i mean there’s definitely still work to be done in this space. in general though inherently i want to be on the side of the cheap cards until proven otherwise, there are a lot fewer natural problems so you have a lot more room to play around with to solve the really problematic stuff.

• i guess i’ll talk a bit more about kuruk ctrl because i am a nonbeliever. afaict the kuruk ctrl deck is just a blue small deck except instead of all its cards costing 1 all its cards cost 2. i mean wrath and elspeth are nice in many cases. but like the thing about the small creature pressure… we have a tool for that, it’s stormchaser’s talent, it even draws a real card for 4 mana instead of a fake card for 5 mana. i mean you could maybe still explore stuff in this space with the white big cards even, i just think it seems better to start with the blue small structure and add stuff to that instead of structuring around the far more inefficient jeskai ctrl cards.

Smol?

• i mean there's also stuff like red ojer or leyline or like green landfall. allies?? i don't really have much faith in this stuff but i mean we should probably not ignore it totally since there are probably some red sickos or something + even think about them as exploit decks if it seems like people are low on the right kind of interaction, at least personally that's like a last resort kind of deck selection mechanism but maybe others feel differently. should try putting modern staple legend of roku in your red deck maybe

Misc.

• i do want to be sure that i don't get totally locked into frameworks of what decks must look like, this is how to miss a deck like pixie that is like Every Horrible Raven's Crime Deck Ever Except This Time It's Actually Broken. idk if e.g. icetill fits that well into the above framework, though it has a lot of molelike and beeglike properties/issues at once i feel. i will also call out this deck https://www.mtggoldfish.com/deck/7471848#paper that i got crushed by while playing izzet discard, though probably that was at least a little due to me not really knowing what was going on. people who like random annoying white creatures could maybe check out that one (and i will note as always that surely a splash in your monocolor deck is pretty free).

There's a lot in here, but if I had to summarize, it would basically be that there were two theoretical pillars of the format: Badgermole Cub and Stormchaser's Talent, with a variety of other decks like Control, Living End, Leyline Red, and Icetill sprinkled in. The Stormchaser's Talent decks ('Blue Small') were theoretically well-positioned into the badgermole decks because they could play lots of cheap interaction and never run out of cards, but it wasn't clear how to build the best version of them, and they struggled to actually close games because they didn't have a good "output engine".

......

......

......

Yeah. cft vindicated once again.

You can see how close we were.

Want to see us missing it in realtime?

Anyways...

That's about as far as we got with the Lessons deck. Cft had noted it as the version of Blue Small with the best card engine out of all of them, and tried the non-monument version a bit, but felt like it was too weak against control without the same amount of selection for its non-dead cards, and too weak to graveyard hate to be worth exploring further. I tried a version with Legend of Kuruk as an output piece that could close games alongside Frostcliff Siege, but concluded that the lessons being slightly less efficient than normal interactive spells, and Kuruk being less efficient than Stock Up, was too much clunk. Jesse played against the Monument Talent version once on ladder and came away feeling like it was extremely powerful, but we didn't take this as a sign that we needed to try that version; we dismissed it as inconsistent without actually playing any games.

You have to discard a lot of potential decks during a process like this, but I think our biggest mistake here was simply not playing any games with a deck that had put up some notable online results and felt strong when we saw it once on ladder. We played at least a few games with almost every other deck that showed up anywhere notable.

Here is a list of marginal decks we spent more time on than Izzet Lessons:

• Izzet Splash Portal

• Izzet Floodcaller

• Dimir Bounce

• Leyline Red

• Innkeeper's Talent Red

• Bant Ouroboroid + Virtue of Loyalty

• GB Ouroboroid

• Living End

• UW Tempo

• GW Collector's Cage

• Allies

Partially, what I think happened was that Izzet Lessons with talent and monument was a subcategory-of-a-subcategory. We actively worked on many Izzet decks, and knew there were likely multiple strong decks in the space (in addition to Izzet Discard, which was clearly strong and winning a lot online, but we beat pretty consistently in our internal testing with a variety of decks). We felt like we had tried "Izzet Lessons", but we hadn't tried Izzet Monument Lessons. And that small decision to dismiss the deck on pure vibes for the sake of time efficiency, the kind of decision you make a many times over for a variety of different decks in the leadup to every single tournament, cost us dearly.

Post-Miss

Okay, so we didn't find the lessons deck. That wasn't actually the end of our process, it was just a blip that felt insignificant at the time. Returning to the actual timeline, we had just had our big pre-house standard meeting, and roughly everyone agreed that the field was divided into a wide variety of badgermole decks, a mysterious and likely-containing-multitudes Izzet space, and some other medium-strong contenders and decks-to-be-aware-of like Jeskai Control, Living End, Dimir, and Icetill.

Where was I at, personally?

Mainly, I was trying to not end up registering Otters.

I haven't yet mentioned that one of the top-contender badgermole decks we had developed was a version of the Airbending combo deck that was heavily played at Worlds (and popular online leading up), but was also playing 4 maindeck Ouroboroids as extra hits off of Aang 5 and a single-card backup "fair" plan in case the opponent had removal to pick apart your combo or you couldn't find the right key pieces. It felt very difficult to beat this deck with anything except a pile of interaction and sweepers, because you never knew which plan they were trying to implement. You'd torch their turn 2 Doc Aurlock and get blown out with an Ouroboroid the following turn, or let the Aurlock live, and die to them setting up a removal-resistant combo.

Lots of hype was starting to form around this deck, so the night before Thanksgiving, I queued up into it on Otters in an attempt to bully myself off of Otters, which had continued to put up strong-but-not-incredibly-strong overall results. For the first few matches it looked like the plan was working to perfection, but then...

I hated that this was the conclusion I was drawing from our testing. I really didn't want to be the person saying "otters is broken you just have to play it perfectly", because that was such a stereotype of Sanctum and it fed into my previously-existing beliefs of Otters actually being an amazing choice for last year's worlds that didn't end up panning out for the rest of the team because of a combination of luck and preparation gaps. I find that the more new things I feel like I'm learning from any testing process, the better things tend to go for me in the tournament, but it felt like all I was learning here was old things.

Otters continued to put up very strong results against the Badgermole-Ouroboroid decks we thought were going to be the most popular decks in the tournament...

I tried to replicate my otters sets against a different pilot on Airbending...

Jason had a somewhat-sketchy set against Izzet Discard (which we believed would be the most popular Izzet deck, and possibly the most popular deck in the room)...

Jesse wanted to try her luck against Otters with Airbending...

I had a slightly concerning set with Otters against Paul playing normal Dimir mid, but learned I'd been failing to recognize some play patterns like Azure Beastbinder not actually taking away your Badgermole land mana for flash creatures like Floodcaller...

Jason tried checking my Dimir results and found things to be Otters-favored with the right piloting...

... and re-re-checked the matchup with normal Simic aggro...

... but cft wasn't convinced, so we checked again:

While all this checking and re-checking was going on, I was waking up at 5am the morning after Thanksgiving and boarding a 7am flight to the testing house all the way across the country. Once we got settled in and I had a chance to get some sleep, we were just days out of needing to submit decklists, so I got to work gauntleting Otters against the last problematic style of deck -- Izzet (all varieties). First, I wanted to try my Izzet Floodcaller deck that had been good against the other Badgermole decks like Simic, to see if a non-aggro Izzet deck gave Otters the same problems. It didn't seem to be the case. However, the Izzet Discard matchup still felt like a problem, at least for me:

At this point, I was slightly worried that a lot of our Otters testing results had, for better or worse, been a result of Jason skill-diffing people with a deck that was only good if you played it perfectly. Despite that though, I still personally felt pretty good about every matchup besides maybe Izzet Discard and Dimir Mid -- but those were two important matchups in my view.

As all of this Otters testing had been happening, another deck had been showing some very promising results as well...

Swapping out reasonable-card Gwen Stacy for the stinky Virtue of Loyalties gave the Kuruk control deck that we'd solidly liked, but not loved, before, the chance to completely flip what had previously been its worst matchup, boomer Dimir. If I was worried about Dimir and Izzet Discard with Otters, this felt like a natural place to turn. As usual, I tried to bully myself off of Otters by playing a set against the deck I was hoping to switch to...

This actually worked! Despite Otters still feeling strong, the fact that the Kuruk deck could put up solid numbers in a matchup that has traditionally been quite Otters-favored (with skilled piloting, of course) was a very impressive showing. The deck had really good vibes, and if it crushed Simic like it was supposed to, seemed to have a pretty incredible matchup spread. In my head, I soft-switched over to registering Kuruk Control, and made a note to test the Simic matchup the next day to confirm it was as good as we assumed it was.

Well.... that's a bummer.

Even though Casey wasn't swayed off of Jeskai from this set, it felt like a big disappointment for me, whose main reason to register a deck full of sweepers would be that it was strong against Badgermole-Ouroboroid decks. These games made me doubt that if the Badgermole decks came prepared with anti-control tools like Sab-Sunen or Ba Sing Se, control would actually be favored against a deck playing far more efficient cards and cards more powerful in a vacuum.

Notably, I probably shifted too far away from Kuruk and back towards Otters after this set, since afterwards, Casey continued to put up strong showings with Kuruk against Dimir Bounce and Bant Airbending, two decks we thought might be real contenders just days out of deck submission.

That being said, I took this set as the final nail in the coffin: I needed to grow up and once more learn to play otters in every matchup.

And so I did.

cft was still skeptical of the Otters vs. Discard matchup, and they and Jason played a lot of sets with the discard deck bringing in pyroclasms to significantly more mixed results

However, this was fine by me -- even if the possible most-popular-deck-in-the-field matchup was just somewhat swingy and even, I now felt quite confident against...

• Simic Aggro

• Bant Airbending

• Jeskai Control lists without Gwen-Shiko

• Normal Dimir

• Dimir Bounce

• Izzet Prowess

• Reasonably confident there wasn't a good red deck out there that crushed Otters.

It felt nearly silly not to register Otters again for Worlds. I'd figured out extremely solid plans against the toughest matchups, and the deck had nearly no matchup problems beyond being hard to pilot -- but that hadn't stopped me before, and I couldn't let it stop me this time.

I played a few matches against living end, concluded that the matchup was horrible unless we registered not one, not two, but it absolutely must be three copies of Soul-Guide Lantern in our board, cut the 3rd pyroclasm I wasn't sure about vs. the Ouroboroid decks for a Torpor Orb that covered both Airbending and Living End, and went through the mapping spreadsheet one more time with Jason to make sure we were on similar pages about everything.

When we finished, it was half-past-midnight.

Jason said "I think you're going to win this tournament."

I said, "I agree."

And I hit submit, went to bed, and looked forward to spending all of Wednesday on Limited, my comfort zone.

I'd worked my ass off to convince myself not to register Otters again. It was too clean. The deck had gotten an incredibly powerful new tool printed in Badgermole Cub, and had its only natural predators banned. It always seemed to have bad matchups on the surface, but the badness of those matchups faded away with the right piloting and the right plan. We'd felt this way about Otters last year at Worlds, and at least going by the results, we'd been wrong. I was worried I'd gotten lucky last year and was overhyping my own skilled piloting as the reason I did so well with this mythically hard-to-play deck, which was actually an incredibly fragile Sanctum pet-deck that we were going to trick ourselves into playing again.

But I'd done the work, and the results were as clear as they could possibly be: Otters was the best choice if I wanted to win the World Championships - which, this year, I believed I was capable of doing.

The only way this could go wrong was if we'd failed to find some hypothetical broken deck that crushed Otters, and other teams had found & brought that deck.

.... Oops.

What Did We Learn?

I've been licking my wounds over the past week and figuring out what to feel about what happened at Worlds.

Funny enough, I didn't actually play against any Izzet Lessons decks in the constructed rounds. I played against:

• Javier Dominguez on Bant Airbending (1-2)

• Alex Friedrichsen on Handshake Otters (0-2)

• Connor Mackenzie on Izzet Discard (0-2)

• Yuuki Ichikawa on Izzet Discard (2-1)

Which, based on our testing, was 3 even-ish matchups and one good one (though, to be fair, it was Javier lol). So I don't really think I can blame my constructed performance on the metagame.

Can I blame it on the draws just not being there?

I have no earthly idea, and I authentically have no interest whatsoever in seeking an answer to this question. I think this is one of the biggest improvements I've made as a player in the past 2 years. It now truly feels to me, intuitively, next-to-impossible to figure out whether I lost because of 'luck' or because of controllable variables, because I can just see so many things that are under my control now.

Here is a non-exhaustive list of potential mistakes I made in the tournament, just the ones I recognized in the moment and still remember over a week later:

• I picked a Lost Days over a Merchant of Many Hats in pack 2 of the draft, when my deck needed more early board presence and Merchant would have been better

• I missed an opportunity to play Island-Island instead of Island-Swamp over the first 2 turns in game 1 of limited, to give myself the optionality to flash in Wan Shi Tong if my opponent had landcycled

• I got too greedy with my Wan Shi Tong and didn't play it on turn 4 to trade + draw a card, instead I played a 2-drop that could only chump to build towards casting it on turn 6 and eating a creature

• On that turn, I left up island-swamp after casting my Kyoshi Battle Fan instead of island-island, and got punished by Dai Li Indoctrination into my hand without any other nonland permanents

• I played overly aggressive against my round 3 draft opponent who was signaling he had good racing tools out of a black-red deck, and even though I had a hand with good racing tools, the rest of my deck wasn't set up to help me in that race and it was probably ill-advised

• I decided to go for a Vitality->Badgermole line against Javier despite the vitality tapping me out before it resolved and not letting me leave up Torch the Tower, which left me dead to his last 2 cards in hand being Aang (to remand the vitality and keep me tapped out) + Appa or dig to Appa

• I forgot to bring in Rals in the otters mirror against Alex despite having tested them the previous day against cft and being convinced that they were good, because I was out of it enough that I just followed the mapping we had on the sheet

• I'm sure I made some poor gameplay judgment calls in the mirror that cost me in hard-to-see ways, because we hadn't prepared sufficiently for the mirror (this is actually I think an underrated flaw in our prep -- I underestimated the possibility that a full other team would register Otters, making it the 2nd-most-popular deck, and I should have played far more mirrors)

• I honestly don't remember much about how my match with Cmack went, because I was pretty out of it at that point and just playing on instinctual fundamentals, which is a problem in itself -- I think I needed to sleep more in the days leading up to the tournament and generally do more chronic pain maintenance stuff, because I physically felt drained and hopeless after losing the first two constructed rounds when I should have been able to lock in.

• Once I'd gotten knocked out, I was able to play much more freely and confidently against Yuuki, which could be the classic "good draws make you feel like you're playing better", but could also have been a real mental thing of feeling like the tournament was doomed after I'd seen how bad the lessons matchup was the night before and letting that get to me, but feeling more free to just play my game after I knew my run was over.

These are the execution mistakes I can look back on and learn from. I think there are also lots of process mistakes from our preparation that, without overcorrecting and assuming every tournament will be like this one going forward, are still good adjustments to make:

• We should have just played games with the monument-talent version of Lessons instead of dismissing it out of hand. I don't see this as a major mistake, but a minor one that ended up being extremely costly. You have to dismiss random decks from being "worth trying" all the time in testing, but there are multiple reasons why monument-talent was worth concretely trying:

• We believed that "Blue small" was structurally well-positioned in the format

• We knew that the primary weakness of these decks was not being able to close games quickly enough

• We knew that the primary weakness of the Lessons version specifically was graveyard hate, and this iteration potentially shored up that weakness in a nice way

• The deck had had some actual results online, making it far less likely to be simply a bad deck

• Jesse had played against it and it had felt extremely powerful

• We should have been more willing to believe that Badgermole-Ouroboroid decks were very attackable, and that other teams would figure this out instead of just registering the "level 0" most powerful strategy

• I don't think this is a takeaway that should be applied to any tournament except Worlds -- but Worlds is the one tournament where nearly all players are preparing in teams, and internal results can look a lot different from external ones. Because we believed it was very possible to beat these decks, we maybe shouldn't have assumed that they'd be easy to register for other teams.

• A similar argument applies for Dimir mid and Izzet Looting -- I believed these decks would see play because they were seeing lots of play online, despite us continuing to demolish them in internal sets. This belief was justified by my PT experiences of people bringing "stock" decks that are extremely exploitable to nearly every tournament, but I failed to realize how much the dynamics change for Worlds when all your opponents are being that extra level more thoughtful about their deck choices.

• I think I committed to a constructed deck a little too late in the process, and left myself rushing to find good plans for every matchup over a day and a half. While I was able to do this, it was draining enough to leave me overly exhausted by the day of the tournament, because I didn't have enough buffer to rest & reset.

• I overreacted to the Kuruk-Shiko-Gwen deck having a not-amazing Simic matchup, when the deck was putting up strong results against basically everything else in our testing and we weren't even sure Simic would be the most popular deck.

• I think this was the best deck Sanctum pilots registered for the tournament, and I wish I hadn't been scared off of the pivot (shoutout to Talia the GOAT). Taking a pause and doing an honest matchup breakdown, while acknowledging that the level of certainty I had that Simic would be the most popular deck was actually pretty low, would have helped this.

So, there were many mistakes. But this isn't surprising! There have also been many mistakes made in all the tournaments I've done well at. It's impossible not to make many mistakes when doing something as challenging as this.

I'm very proud of my work for Worlds this year. There were a lot of things I did well. I played a lot of serious, focused testing matches. I changed my views multiple times, learned about things I was doing wrong, and corrected them. I think we got a lot more out of the smaller amount of time we spent on limited work than the amount we've gotten in the past even with more time spent.

There are always mistakes, you always need to learn from them, and there are always successes, and you should always try to recognize them and replicate them. This isn't some trivial cliche to make me feel better -- it's part of keeping myself and my process sane, of staying on the right path, of operating as a flawed human trying to optimize their performance in a competitive and complex system.

Sometimes the mistakes end up being more important than the successes, and vice versa.

Sometimes another team finds the broken deck and you don't, and sometimes you miss the mark on metagame prediction in a small, swingy field.

Sometimes tournaments go badly, and while it's impossible and totally useless to say whether this particular one went badly because I got unlucky in the games, it's certainly true that I've gotten incredibly lucky that the first professional tournament to go badly for me ever was also the first one that had no significant bearing on my future qualifications.

Here are some things I said in the leadup to Worlds that I now cringe at:

"It's easy to bring a good constructed deck to every event, you just have to try really really really hard not to bring a bad one"

"There is just a very high correlation between how hard you prepare for any Magic tournament and your results in that tournament"

(I don't think this one's exactly wrong, but I do think it was said far too cavalierly -- *how** you prepare is equally important, just trying really hard won't be enough).*

I was getting too high on my own supply. This game is brutally hard, and no amount of effort can guarantee success - it can only make it more likely to an uncertain degree. Going on a run like I have over the past year can easily make you forget that. I hadn't forgotten it intellectually, but I'd forgotten it emotionally.

Can I let you in on a secret? I was, on some level, waiting for this moment. It didn't feel real that I'd done as well as I had over the past year. Something was off. Magic tournaments don't always go well.

In some special, odd way, all of it, including the successes, feels more real now.

I've tried my absolute best to win at the highest level the game has to offer -- and I've now failed at that goal.

It no longer feels like I'm getting away with something, because I'm not, and never was.

I'm just another player trying to win, and scared they'll lose.

Nothing is guaranteed. Anything worth doing is hard.

Succeeding in Magic is HARD. But for now, I get to keep trying.

Oh, and I get to try with some amazing people, and that part's pretty fun too.