ROI is (mostly) irrelevant in poker

[Greg (FossilMan)]

There was a remark in a very recent thread that got me onto this rant again. From eastbay in his thread titled "Grr. PP Multi-table Tourney Hand":

[ QUOTE ]
This idea that "you can always just play again" is a little specious, I think. The idea is maximum ROI, so when going after your R, you have to worry about your I.

[/ QUOTE ]

EV matters. Variance matters. Bankroll matters. Return on investment is irrelevant. When you're playing poker, the large majority of your bankroll is NOT working. Most of it is sitting in your pocket, your lockbox, or your bank account, waiting there in case you lose tonight and need to go get it. Thus, who cares how much return you're getting on your current poker investment (that is, the game you're in right now)? Would you rather buy into a 1-3 stud game for $100, earn an expected $5 per hour for 5% ROI; or, would you rather buy into the 10-20 game for $500, earn an expected $20 over the same hour, and only have a 4% ROI?

If your bankroll can afford the variance, you generally want to play in the game that offers you the highest EV per hour. At least in my opinion.

Same thing for tourneys and tourney chips. If the play is one that offers +EV, then who cares about ROI? If a play is going to bust you out when you lose, that's poker. It doesn't become correct to fold just because you will survive longer in this event (without cashing), and thereby put off investing into another tourney right away.

Even if ROI is what matters, you need R to make this number positive, and folding when you've got the best of it does not lead to long-term R.

Later, Greg Raymer (FossilMan)