Luck and Skill in 40k - Part 2

Click here to return to Lecture Series articles

So, about a year ago, I threw up a post on dakka that wound up being coalesced into this article. The abstract of this article (for TLDR types), is that what player skill in 40k is is the ability to play exactly the odds that you want. You bring meltaguns in your list because you want to increase the odds of you being able to kill a land raider from 0% up to something, for example. Likewise, the whole point of moving is to get exactly the weapons in to exactly the right ranges at exactly the right targets. The article then goes on to note that because the increase in player skill becomes more difficult the more skilled you get (it's tougher to play a 93.555% than a 90%), and the more your player skill increases, the less it matters (because playing odds from 50% to 90% is much bigger of a deal than playing from 50% to 50.00000001%), therefore, the better you get, the less player skill matters, especially when you consider that it's relative to your opponent's skill. Eventually, as these controlled variables become more, well, controlled, the uncontrollable variables in the game (like the random results of die rolls) matter relatively more. TLDR - the better you and everyone else gets, the more the results of the game are determined by die rolls.

But this post isn't just to rehash this theory. The reason that this is part 2 is because of a conversation I just had with another knowledgeable 40k player. The purpose of this is to get more into detail about what, in fact, skill actually does in a game of 40k.

To begin with, let's assume, for the sake of the platonic, that in any given situation, there are perfect odds to play. For example, let's say that it's the bottom of turn 7, and your opponent has tank shocked a land raider onto your objective. If you kill the raider, you take back the objective and win the game, if you don't you lose, or draw (depending). In this case, the land raider really, really, really needs to be dead. You should play the odds as short as you possibly can to make sure that raider dies. For another example, let's say that it's the bottom of 7, and there's a unit that has no killing power left (say, an immobilized, weaponless dread), and is nowhere near an objective. You can play the odds here as long as you want, because it doesn't matter, because if you fail, your failure won't have any impact on the game. Of course, most odds fall somewhere in the middle, where some things need to be killed more than others in kind of a big gradient.

Now, whatever those odds are in any given situation, let's say that there is a certain odds that are perfect, and that any player should play. If you play the odds too long, choosing to not apply enough killing power so that you can focus on other stuff, you are going to be too likely to fail, and you should have focused on that unit more (oh, I thought those two meltaguns would be enough. Hmm...). If you play the odds too short, you have the opportunity costs of overkill (yes, that rhino is certainly dead, but you moved all your meltaguns over to shoot it, now say hello to my blood talon dreadnought over here...)

So, the question then becomes, what sets the "right" odds you should play for any given circumstance? The answer, of course, is your opponent. Your opponent moved things the turn before, and so changed what needs to be attacked at what level of priority in your turn. If player skill is the ability to play exactly the odds you want, then player skill vis. a vis. your opponent is to give your opponent odds that he will have difficulty playing.

For example, lets say that in your movement phase you move something to make it a bigger threat, and move something else to make it a lesser threat. In this case, you've given your opponent new odds he needs to play (in order to play the platonically "correct" odds). If your opponent has less skill than you, you will be forcing him to play more of the "wrong" odds than he does of you. For example, you may make things slightly more of a threat, and your player's opponent skill might miscalculate the odds as being a much bigger threat (and thus play too short against it), or may fail to perceive that something became more threatening (and thus ignore it, or keep the same amount of focus on it despite it becoming more dangerous).

Of course, this runs into the problem expressed in part 1. The better you get, the more you're going to be playing for thinner and thinner margins, and it won't actually matter as much. Tricking your opponent into caring only 50% about something when he SHOULD have cared 50.00001% about something is scarcely a difference. It's scarcely a difference, of course, compared to the dice. The problem here is that we're talking about 6-sided dice which is a very coarse random mechanic. Tiny changes in the odds out to many decimal places hardly matter when your unit has a 2/3rd chance of hitting and a 1/3rd chance of doing absolutely nothing. Likewise, there is a great deal of coarseness built into the game itself. If you have 5 meltaguns, you can't fire 3.45 of them at one target and 2.55 of them at another. You have a coarse, discreet set to deal with. Like part 1 notes, as player skill improves, it breaks past this threshold where the deviation becomes more important than the average. If the slightly smarter odds is to play assuming there's a turn 6 and there isn't because you rolled in the bottom 33% of possible outcomes, well that's just that. The game is done, and the "correct" play saw you lose the game. Perhaps these changes will allow you to win more in the very long term, but the smaller the differences, the longer this term becomes (especially compared to, say, a tournament in which you only play a small handful of games).

But there's another way to look at this relative skill gap. One way is to get your opponent to play the odds sufficiently fine when he otherwise has the ability to play them, but another way is to prevent your opponent by playing the right odds by making him less capable of doing so. For example, if I have a bunch of tanks, and I kill off all my opponent's anti-tank, then he is extremely reduced in his ability to play certain odds. If I destroy my opponent's ability to hurt a land raider, and I tank shock his objective turn 7, my opponent literally cannot play the odds required to win. No matter how skilled he is in the movement phase, he can't use dead units.

More importantly, this explains why 40k is so dependent on list building. The tools that you bring determine the kinds of odds you can play. Be deficient in something, and your opponent can always set odds that you will have a very difficult time using player skill to match, if you're even able to at all. This also does a good job explaining why spam lists have become so popular. In the case of a foot horde, you may well bring more models than they brought bullets. No amount of movement is going to help if they don't have the tools at their disposal to be able to play the odds that you've set impossibly high for them with your spam list.

So what skill boils down to, then, is bringing the right list and moving things in such a way where you exactly play the correct odds that your opponent has set for you, and to do it in such a way where you set the odds that are the most difficult for your opponent to play against. Think of it a little bit like tennis. The point is to get to wherever your opponent hit the ball to, and then to hit it back to the place your opponent is least able to get to in return.

The real problem here, though, is that 40k isn't tennis. Tennis works by your skill in running, and your skill in hitting the ball where you want it to go. 40k works by your ability to play and set odds, but it's still setting odds. Just because you played and set the odds of something as exactingly as you possibly can in no way guarantees that you're going to be successful. Only the dice determine that. You can never know if you played the odds wrong and got your just deserts when you failed, or if you played the right odds and the dice failed you (and perhaps if you'd just brought that one extra meltagun, you would have actually been playing worse odds (too conservatively)). Likewise, as mentioned before, making an improvement that's 1/100th of something better will only win you 1 extra game for every 100 you play, and the tiny margins are utterly lost on any particular game because of the coarseness of the system itself.

And this, of course, brings us to the tennis analogy again. If you are 1/100th better than your otherwise exactly equal opponent, then you will volley back and forth 100 times before the better player finally wins, while if you are twice as good, you probably only need to hit it back once or twice for you to win. But the transmission system for skill in 40k is very different than this. Because skill only competes against skill through the medium of luck, it is the luck that is important once you reach the coarseness of the system. In a game of tennis where player skill was filtered through who won a coin toss, instead of playing a game where each game took 100 volleys, but the better person one every game, instead you have a game where each player wins half of the games, with a slight bias towards the person with the relatively higher player skill (which, I should mention once again, decreases as the players get absolutely better). In this case, rather than it being 100-0, it would be 51-49.

And so the competition in 40k is really trying to shake the other person into playing the wrong odds, but, of course, you can never have retroactive confirmation of if they played the right odds or not, because it's not the mere playing of the odds that determines the outcome, but rather the outcome of the die itself. You can never know if your opponent brought one too few meltaguns than they should have if one of the meltaguns blows up your tank anyways. The problem, then is that player skill is actually hidden from sight of human eyes behind a veil of randomness that we have difficulty understanding. It makes it easy to fail to see how lucky someone was, and then make a statement about their player skill based on if they were successful, when the determiner of success was, in fact, dice. In fact, it's possible to come to completely wrong conclusions. You see someone fail to kill a tank and you come to the conclusion that they weren't playing conservatively enough, when it was possible that they were playing TOO conservative, but it just so happened that their dice failed them imperceptibly more than they should have.

So, if player skill is playing odds, and the odds obscure the player skill, then the only way that you can see player skill is over a large sample size. The better you and your opponent get, and the smaller and more refined of odds you play, the larger and larger the data set has to get for you to be able to see the impact. Between two very good players that are able to play odds down to the 10,000ths, you may well need to have them play 10,000 games to see who is the better player. Of course, because this is also relative skill, in this case you would need to play 100,000,000 games to actually see the difference because you're really trying to see where one player played the 10,000ths and the other player only managed a 9,999.

As such, the point at which player skill fades into the sample size is shockingly quickly. For example, if you play three games against someone, they could easily be three times worse than you and still pull out an overall draw. Things are just too coarse, both in data creation (6-sided random events) and in measurement (inability to capture fine changes without lots of data points) to be able to measure the difference in player skill.

I think this also gives a way of thinking about baseline competency in 40k. Once you are able to play the odds to within the coarseness of the system itself, then increases in player skill absolutely pass into the realm of luck (and relative difference in player skill long before then), which would count as mastery in my book. In any case, any effect your improvement you make in your skill becomes increasingly tiny as the improvements themselves become absolutely and relatively tinier. Perhaps that's the point at which you stop caring and just start playing 40k as the dice game that it really is, as any difference in player skill may show up only once a year, or once every decade, or, practically speaking, once every never.