Sign Up For Card Player's Newsletter And Free Bi-Monthly Online Magazine


Poker Training

Newsletter and Magazine

Sign Up

Find Your Local

Card Room



by David Downing |  Published: Jul 01, 2008


Sometimes, rarely, every now and again, a whole group of people realise they are wrong; barking up the wrong tree; holding the wrong end of the stick. The rather grandiose term for this is paradigm shift. I prefer a different analogy, which was once made about science; scientists can be viewed as all chipping and digging away at a pit, happy and dirty. But occasionally they realise they are digging in the wrong place. The hole should be over there. Poker, as a group of theories, has not really got digging yet; however, I like to think we are doing our level best to point out when players are digging in soft, grainy sand -- and the tide is coming in.

One of my favourite suspect excavations is the idea that 300 big blinds is sufficient enough a bankroll for a winning limit hold'em player. The problem here is that very few advocates of this theory actually know how it was derived. The first bombshell comes with one of its first assumptions; a hand of poker can be treated the same as a hand of blackjack or a coin toss. Hopefully, you're saying, "What!?!" just as I did when I first read the proof. This is a hell of an assumption to make, with no real empirical basis except that it makes the calculations easier. The second weakness is that it assumes poker results follow a certain statistical model called the "normal distribution." The arguments against this are better suited to Mathematics Monthly than Card Player, but trust me, there are serious concerns that this is really the case at all. Lastly, and far more damning, is that the original theory was solely concerned with starting-bankroll requirements, and not what your ongoing roll might be. So, if you continually make "withdrawals" once you get more than 300 big blinds, you are not following the precepts of the theory at all. The warm, fuzzy glow of knowing you are mathematically safe is just warm and fuzzy, with all of the already suspect mathematical certainty removed.

Another hole that needs filling is the obscure ideas people have about variance. The conversation often starts off along the lines of, "Does game X have less or more variance than game Y?" normally comparing some version of hold'em to some version of Omaha. And by variance, they really mean, how many and how painful will my downswings be on my pleasingly upward-pointing poker-results graph?

I would argue that the game itself is the least of your worries.

Obviously, the more cards you get to play, the potentially thinner edges you will find yourself having to pursue, so the amount of variance will increase. One-card, no-ante, no-draw poker is almost variance free. It is also not played. Five-card stud high was its nearest equivalent, hence its current popularity. Some of the "it's all in the game" confusion is added to by the great and the good proclaiming that the split-pot element reduces variance. I'm sure this was the case back in the old live games when high-low Omaha or stud was populated either by nits or very solid experts. But the games online often do not resemble this at all, with "Grumpy Old Men" being replaced by "Manic Teenage Donks" with the games correspondingly transformed.

This leads me into a factor that I believe at least equals, if not outweighs, game type: the opposition. For a good chunk of last year, I played $30-$60 stud eight-or-better and its Omaha equivalent on PokerStars. These games were far, far looser than their lower-stakes equivalents. The challenge then becomes whether to nut up -- play tighter -- or man up -- get in there and start fighting. For example, if you know that your foe will bet up to and including the river with no pair and no low but a better-looking board, you either play a waiting game or start calling on the river with just a pair of deuces. Either way, if the "loosies" are not completely insane, your variance is likely to increase. In the first case, not by as much, but your win rate may be at risk; in the second, by much more, but hopefully with a corresponding increase in your average win rate.

So, what is the key factor in the variance equation? Unsurprisingly, it's you, of course. It is perfectly possible to have a losing player who bizarrely has very low variance. Perhaps surprisingly, this type is more common than you might think. A very tight, passive, but dumb, unthinking player might gently leak small losses almost forever, without experiencing either highs or lows. Juxtaposed to this, a break-even player could have enormous swings and a nightmarish variance. One way of "achieving" this would be to have a serious tilt problem. Your winning game may be a profound A+, but if your tilt game starts to approach the tail end of the alphabet and metaphysically beyond, these twin states could easily cancel each other out; zero profit, maximum pain. Winning players could easily have both extremes, with a rock-like, multitable player having minor fluctuations compared to his very loose, aggressive, but highly skilled tablemate.

It is arguable that as long as win rates and bankroll management are good, variance is an irrelevance. In fact, a high-variance game is something to be relished. If both you and the game are capable of experiencing substantial swings, then (a) you are likely to get action from other players, (b) bad players will have bigger and better runs of good luck and will survive longer, and (c) as an outcome of (a) and (b) and in a virtuous circle, feeding both, your game will attract loose, gambling players.

A winning player needs to clamber out of the trenches and decide for himself how he will affect the game, and how he can maximise these benefits and not arbitrarily allow the game to affect him.

David has played poker all over the UK for the better part of a decade. Originally a tournament player, now focused on cash play and almost entirely on the Internet for the last three years, he makes a healthy second income playing a wide range of games.