Tuesday, September 20, 2005

Cold Calling: The Silent Killer

One of the obvious benefits of data mining the hell out of Party tables lately is that you end up with all sorts of nice, juicy data, above and beyond your own personal stats. I've spent a goodly amount of time lately reverse-engineering some of it, as far as finding the most profitable players and then working backwards, seeing if there are any obvious tendencies or trends.

The interesting thing is that while you can come to some pretty iron-clad generalizations (e.g. you can't be a winning player with a VPIP over 40%), there's a surprising amount of divergence in many of the indicators that people focus on. It's hard to hang your hat on any one number, circle it, and say that's where you should be. If you look at enough data, you can find all sorts of exceptions to many rules you'd like to create, as far as what AF or VPIP a winning player should have. In general, yes, we can arrive at an average that serves as a very good benchmark to shoot for, but my point is that there are plenty of statistical outliers for nearly every factor you might isolate, with solid sample sizes behind them, that lead to a bit of head-scratchin' when you start examining them.

The one factor, though, that I've been able to observe that has the best correlation to profitability of play, is the percentage of time a player cold-calls. The most profitable players cold-call the least, with the percentage of cold-calls increasing slightly as you move down the profitability range.

Again, makes perfect sense, as we've all learned (hopefully) from all sorts of poker authors that you almost never want to be calling two or three bets. If you have a powerful hand, raise. If it's not powerful enough to raise, then fold, as at least one other person is displaying strength.

The problem, though, is at heart we're all monkeys that like to play poker. So while we may logically know that cold-calling gives the baby Jebus peptic ulcers, it's really, really hard to toss away that A10d, when it's raised and re-raised in front of us. Especially if that's the best hand we've seen in 20 minutes. It's much easier to just call, crossing our fingers and chanting "Diamonds diamonds diamonds diamonds".

An additional problem is that we have very selective minds, which are prone to remember the isolated incidents where we cold-called three bets with A10d and flopped the nut straight, and immediately expunging from our memory all the times where we cold-called and whiffed (or, even worse, flop an A, can't get away from it due to the size of the pot, only to end up out-kicked and substantially poorer.)

Again, I'm not trying to re-invent the wheel here, as this is relatively common-sensical and widely broadcast by all sorts of poker authorities. What surprised me, though, was not only just how accurate an indicator cold-calling frequency was, as far as profitability, but just how infrequently the most profitable players cold-call. One of the more profitable players in my database has cold-called exactly 6 times over 44,000 hands. It's not uncommon at all to see cold-call percentages of less than .1 percent among top players.

While on the surface that seems discouraging, as far as advocating you throw more hands away, in the face of a raise, it's really not. If anything, it adds support to the idea that aggressive poker is the way to roll. If you're going to play that pair of tens, then three-bet, instead of meekly calling two bets cold. You'll only flop the set 1 in 8 times, so you have to maximize what you win when you do for it to be profitable to play those tens. Ditto for that A10d.

Like many pokery decisions, the expectation between folding and raising with good but not great hands (when faced with a cold-call decision) is usually often pretty small. The big pots you take down when you hit at best barely offset the times you completely miss on the flop and have to fold.

What can kill your long-term profitability, though, are the cold-calls. That's the worst of both worlds, as you invest at least 1 BB in a situation where you're likely already behind, yet you do nothing to maximize the pot when your hand is ultimately good. You're voluntarily risking additional money in a marginal situation yet also refusing to increase your potential profits, in the relatively rare instances that you win.

One thing to keep in mind, too, is that being in the blinds doesn't mean you're not cold-calling. This has been a bad leak of mine that I'm working to address lately, especially when in the SB. Say you're in the SB in a 15/30 game and it folds around to the button, who's been stealing liberally with just about anything. Button raises. You've got Q10c. Since you're opening up your requirements to defend, given the circumstances, you're pretty happy with your hand. You call, waiting for the BB to fold and get out of the way, so that you can go to war with the button, ready to check-raise the flop and really putting Mr. Stealy McStealsaLot to the test.

Bzzzt, wrong answer. Because of the raise, you're essentially cold-calling, even though you're in the SB and have already invested $10 in blinds. Don't delude yourself into thinking this is different from the horrible, evil cold-calling that you've been warned against. Slightly different, but it's basically the same, just a wee bit less expensive than normal.

What's worse, though, about the cold-call from the SB is that you have a chance to knock the BB out of the hand and pick up dead blind money for free, thus making a raise in that situation even more profitable than normal. While meekly completing from the SB is correct in many, many cases, especially if there are mucho limpers, I can't think of any situation where merely calling a button open-raise from the SB is correct. Aside from all of that, you're also likely taking the momentum back in the hand with a raise, especially if the button is on a complete steal, and can take it down on the flop without much of a fight.

To recap: cold-calling gives the baby Jebus peptic ulcers. Don't put him through that. He's suffered enough for your sins. Either fold or raise.

8 comments:

Todd said...

I think you've hit on the converse of an essential poker truth. You looked for trends of consistent winners to copy them. I think it's more effective to look for trends from losing players and NOT copy them. Plugging leaks is better than finding gimmicky plays ("stop and go", value betting, leverage betting, etc.).

p.s. You're goin' down this week! Bow to the power of the Sooted Connectors!

total said...

Great post, just wish I could have read that before going down $15 playing .5/1.00 limit over at Party. I wish I had know about poor baby Jebus, If I had, maybe I woulnd't have gone down that much. Oh well, live and learn.

Unknown said...

Biggest leak in my NLHE game, cold calling with TT, JJ, AK. Bad bad bad.

Easycure said...

I can't wait to get home and see this cold-calling leak in my game. Thanks for making me better.

GaryC said...

"especially when it's the best hand you've seen in twenty minutes"

Truer words were never spoken.

Great post.

G

Gnome said...

Excellent post. I wonder if cold calling may be appropriate with low pocket pairs in late position behind a raiser and four or more callers? What do you think? If so, could the same logic be applied to small suited connectors?

Chip said...

Great post. Plugging leaks is essential and cold calling is probably the biggest leak in most breakeven players.

Dave said...

I've started exporting poker tracker stats up to my Poker Stars client and now track how often a player cold calls in games. If they are a losing player and they cold call more then 2% of the time then they get on my buddy list.

It's difficult to want to throw away hands like TT, JJ, AKo. It's a matter of chosing the right situations, especially in a tourney or SNG.