Uncategorized

Methods and Madness: Analyze This!

I’m amazed how well you handle defeat, for a rookie.
– The Bronx, “Young Blood”

Going back to school so late in life has been a rough ride. It’s tough to appreciate how much better working for a living is than the alternative – no matter how shitty your job might be – until you’re living on eight hundred bucks a month when you’re used to three to five times that (thanks, oil, I will miss your epic paycheques!). Still, better than nothing I suppose, and there is the added bonus of the great big career-shaped carrot & stick sort of thing in play here as well. Between fighting off poverty and time commitments to my ed-you-ma-cay-shun, my card-slinging has been limited to nearly zero these last two years. Well, except for the trip to Montreal in July that nearly broke my spirit for good, and my first 0-2 drop PTQ in ages not long ago in Calgary. That’s just what happens when you don’t practice or put in the work, and I’m over it. Whatevs.

With all of that out of the way, I’m looking to get back into the groove re: Magical cards. As a result, I’ve been thinking about some of the reasons I’m able to do so well at other things I set my mind to, and how I can apply those same techniques to my lifelong hobby-slash-obsession. I’d like to share some of my conclusions today, and maybe there’s enough here that others in the same boat that I’m in can take something valuable away from it all by the end. There may also be history lessons, oblique drug references, and one (maybe two) decklists – but I promise nothing.

The process of belief is an elixir when you’re weak.
I must admit at times, I indulge it on the sneak…
– Bad Religion, “Materialist”

One of the things that really gets me going is planning. Plotting, scheming, conniving, conspiring, colluding – these are all things I have done, and enjoy doing. At least in the abstract, I am a schemer, as are many of the gamers I’ve known to this point in my life. We have to be, right? Every phase of every turn of every game in every tournament is a plan being executed, one step at a time. The plan is constantly changing, with constraints dictated by circumstance. People that just run wild through games of Magic are the ones that don’t win much. I think one of the most important things we can do to improve as players is to fully embrace this aspect of our character.

This is something I cannot stress enough, but I’ll try:

DEVELOP A PROCESS FOR PLAYTESTING

That’s probably not enough emphasis, but I don’t know if I’m allowed to put it in flashing, marquee-scrolling, straight out of circa 1995 HTML hell 144 point Verdana font, so it’ll have to do.

I’m not kidding though, this is important. Too many times around here (read: Red Deer, Alberta. Yeah.), I see people who are quote-unquote playtesting with no records kept, no notes taken, no sideboarded games played, insufficient time with each deck to learn it’s ins and outs – I could go on, but my team of physicians tells me that stress is my mortal enemy. If you want to generate useful data from testing, it doesn’t take a hell of a lot of work. A piece of note paper with tally marks is a good start. If you’re feeling a little more ambitious (and I hope you are), you might even go as far as to – gasp – take notes. My favorite tool in playtesting, though, is spreadsheets.

I have a couple that are patently absurd, actually. With enough time, you can build a spreadsheet that, given the right sample data, will tell you things like highest average match win percentage in a metagame, winningest post-sideboard deck, and who’s game one win ratio has the strongest statistical correlation to winning the die roll. You can also weight your data so that wins against your predicted most-popular deck matter more than others, which is useful for attacking skewed metagames . Right now (in Standard –maybe  Extended, too) that deck would almost certainly be something with equipment, [card]Stoneforge Mystic[/card] and [card]Jace, the Mind Sculptor[/card] and either a Black or a Red splash. Cough cough. Am I going to tell you exactly how to craft such a tool? Well, no. And I’m not going to hand mine over, either. Not because I want to keep it a secret, but because the development of a weapon like that should be a part of your process, too. Understanding odds, percentages, simple statistics – these are all things that will make you better Magic players simply by exercising your brain.

Gimme, gimmie, gimmie. I need some more
– Black Flag, “Gimmie gimmie gimmie”

Maximizing the utility of available data definitely pays dividends. I qualified for Nationals in 2010 with a Naya aggro deck that I picked up out of a Brad Nelson article the day before the event. However, I was very comfortable switching to that list because we knew from methodical and honest testing that our “bad” Naya deck was performing well enough to at least be considered for that tournament. When I saw the improvements over my version in the list from that article, I knew I was playing it. Turned out to be a good call.

[deck title=Naya Vengevine – 2010]
[Lands]
2 Raging Ravine
2 Stirring Wildwood
2 Misty Rainforest
1 Verdant Catacombs
4 Arid Mesa
1 Sejiri Steppe
5 Forest
2 Mountain
2 Plains
1 Tectonic Edge
2 Terramorphic Expanse
[/Lands]
[Spells]
3 Ranger of Eos
3 Oblivion Ring
2 Path to Exile
4 Lightning Bolt
1 Goblin Bushwhacker
3 Vengevine
4 Wild Nacatl
1 Scute Mob
4 Noble Hierarch
2 Birds of Paradise
4 Knight of the Reliquary
4 Bloodbraid Elf
1 Behemoth Sledge
[/Spells]
[Sideboard]
2 Stoneforge Mystic
3 Manabarbs
3 Dauntless Escort
4 Cunning Sparkmage
1 Basilisk Collar
2 Qasali Pridemage
[/Sideboard]
[/deck]

Looking back on it, this deck was just terrible, too. The Sparkmages should have been in the maindeck, Birds sucked, only 3 [card]Vengevine[/card]s…wow. But it worked. And yeah, I did all of this work but ended up copying a different deck at the 11th hour. So what? I was making an informed decision based on pre-existing knowledge of the format, and that’s what you get from building your process.

So of course, in the run up to Nationals itself, I abandoned my careful preparatory process in favour of just running Naya again. Not the same list, obviously, as Gerry Thompson and others had turned Naya into a very different deck by the time Nationals rolled around, but the principle (attack for 4, cast [card]Bloodbraid Elf[/card], rinse + repeat) remained the same. For those who remember, our Nationals and the others that weekend were basically the launch party for Valakut (Hi, Vincent!), and my poor little pet deck just couldn’t rumble with the new kid on the block. Nationals 2010 raped my ego, and it was all for lack of following the procedures that got me there in the first place. Don’t let it happen to you!

Look in the mirror – it’s not enough.
What makes you think that you’re not one of us?
– Face to Face, “Ordinary”

It feels like I haven’t talked enough about exactly what I mean by a “process” for playtesting. It’s a tricky subject, because what works for me really might not work for others. The point is that you have to approach it with a plan. Jamming infinite games of Random Deck v. Random Deck helps nobody. On the other side of the coin, extensively practicing every single matchup in a given format is a waste of your time. You need to know what shortcuts you can take to reduce the number of variables – i.e. decks – being studied so that your testing sessions are productive. Here – for reference/ridicule – is the way that I prefer to approach a format that is already underway.

Look at recent top 8’s, in real life and on Magic: Online. Tabulate those results and take, say, the top 4 or 5 performing decks overall and see how they do against one another in a decent set of test games, but don’t go overboard at this stage. For the first pass, you can probably get away with 4 pre-sideboard games and 6 post-board games, to give you a sense of where the important sticking points are in a given matchup. Even in a tiny 5 deck format, testing every matchup this way is 100 games in total – more decks increases this number rather quickly, so discriminate with great prejudice and look at only the top decks to begin with. I should also note that it is absolutely critical that you play twice as many sideboarded games as game ones. Doing so here skews your game win percentage statistics later on to favor sideboarded spell-slinging, which is exactly how things are in a tournament setting. Record your results, and make sure to pay special attention to weird interactions, interesting board positions, or anything at all that seems important.

With that done, you can take another look at the big picture. First, who won the greatest percentage of games overall? Pre-sideboard? Post-sideboard? Look for things like a high percentage of pre-sideboard wins alongside a low post-board record. A deck with those numbers is either easy to hate out, utilizing a bad sideboard, or both. On the other hand, a deck with low game one percentages and high game two and three stats may have a great sideboard, but seems to have a weak Plan A against the field. I have to stress that these observations are very much preliminary ones, and that it is possible with a small sample size to have skewed numbers due to variance, acts of God, or whatever. At this point, you’re looking for the beginnings of a pattern, and a reasonable starting point for further investigation. A final caveat: thus far, I’m assuming a metagame in which all of the candidate decks will be equally represented. This will never be true for a given tournament, but it’s up to you to weight your results in a way that takes this into account. A good rule of thumb is to assume that everyone else knows what you know about a format, and that the observed “best” decks – the top two, for instance – will be played more than the others.

Once this first pass phase is complete, and you’ve tabulated all of your data, pick the deck that performed best. Keep in mind that what performs best is up to you to decide, but generally involves having a high overall win percentage against your gauntlet as well as a sideboard win percentage that is greater than or equal to its game one win percentage. In addition, look for a deck that is not only beating the field, but also tending to beat the next-best deck more often than not. Finally, absolutely avoid decks that tend to lose post-board unless you have a very compelling reason not to. What constitutes a compelling reason? In the hands of the experienced, extended Dredge circa 2007-2008 had an enormous game one win percentage, but once people started sideboarding in four-ofs like [card]Leyline of the Void[/card], [card]Tormod’s Crypt[/card], [card]Extirpate[/card], or a combination of all three (and more, for the completely paranoid), you had to use most of your sideboard to battle theirs. What this meant was that for awhile, Dredge would win game one close to 100% of the time (yes, really), and game two or three about 50% of the time. This would be one of those situations where my final rule could certainly be broken.

Having selected a deck, you need to go back over your data and find what its worst matchups are, looking at the costs and benefits of the kinds of changes required in order to shore them up. This is where taking lots of notes during games will pay off. Knowing what matters is a huge part of constructed Magic, and no detail is too small. For instance, in a Standard environment where the de-facto removal spell is [card]Go for the Throat[/card], having threats that don’t die to it such as [card]Wurmcoil Engine[/card] is a big deal. Obvious statement is obvious, but you can see where I’m going with this. Pay careful attention to how you win and (more importantly) lose during playtest sessions; it will pay huge dividends in the long run.

Now that (presumably) some changes have been made, run your deck against the gauntlet again, but this time increase the total number of games you’re playing. Since you’re only testing one deck against the field, you can devote more time to specific matchups. In our imaginary five deck metagame, you could now put in 8 pre-board games and 16 post-board games against the other four decks for approximately the same total of 100 games played overall. Now your numbers start to take on a greater  statistical significance, as more iterations reduce the variance you may have experienced in the first phase with its small sample sizes. After playing sufficient games to satisfy your own requirements for completeness, compare your new set of results with the first pass. What changed? Are your win percentages the same, better or worse? Did you notice anything new that made a big difference in performance? Did the changes you made work, and what did they cost you in your “good” matchups?

I think it goes without saying that this process could theoretically continue into infinity. At some point, though, you’ll run out of time and have to go actually play in a tournament. If no huge changes occurred between the starting point and your tournament (such as the release of a new set in Standard or Extended), you should have a very thorough understanding of not only your chosen deck, but the format as a whole. Knowing what’s going on across the table from you is another huge advantage, and it’s a pleasant side-effect of a methodical approach to constructed preparation.

My wordcount is ticking up higher than Charlie Sheen’s Tiger Blood-augmented heart rate, but before I run out of steam here, let me restate the point: I don’t care if you think my way of preparing for a constructed tournament is brilliant, amusingly wrong, or just plain dumb. What matters is that you recognize the importance of your process – dirty, unpleasant, and sometimes downright joyless though it may be – as you navigate to a better understanding of the major players in a constructed format. Along the way, if you can build an accurate statistical representation of the format, you will be well-equipped to adapt to new and changing technology as it emerges. Assuming that you’re not naturally gifted (which is probably the case, if you’re not a Pro Tour regular) enough to just “get” formats, decks, and everything else I’ve talked about here, adopting a methodical, analytical approach to constructed Magic will help you make up for it.

Subscribe
Notify of
guest

0 Comments
Inline Feedbacks
View all comments