Time to burst the bubble?

Time to burst the bubble?

Friday, 15 February 2013

There was a time when one of the biggest opportunities to exploit opponents in poker was on the bubble of a tournament. Players who had satellited in, wanted to get a flag for their Hendon Mob, or were otherwise just desperate to cash – literally at almost any cost – would fold everything to ladder their way into the money.

They had read once in Sklansky that it was rational to fold aces and that had a nice ring to it. All they had to do was avoid trouble. Often such players would openly declare this strategy. And there was always someone who piped up about a bubble prize. For professionals this situation was a turkey shoot.

But times have changed. I can’t remember the last time everyone agreed on taking money off the top prizes for a bubble consolation. There’s always (at least) one stoical internet kid who blocks it, regardless of the protestations from the part-timers and the spin-up merchants.

More significantly, seemingly everyone knows correct bubble strategy these days. In any tournament with more than a £150 buy-in, bubbles are now turgid affairs, endless folding punctuated by the odd open-raise and take down. I must confess I don’t know why the internet kids are so resolutely against creaming off a bubble prize: it’s not like I’ve seen any of them hero-jam on a poor craven soul who has dared to get involved in a pot. They just seem to be folding like everyone else. So much for exploitation!
Don’t get me wrong, there are still spots to be exploited. But is it anywhere near like the good old days when the bubble was basically free money? In Vegas, maybe. In live satellites into festival main events, possibly. But as a general rule, no.

If that’s the case, we need to have a bit of a rethink about bubble dynamics. I raise this issue now because I found myself on the bubble of a £500 NLHE event at the Vic last month which lasted two and a half hours. The crazy thing was that I didn’t even rate many people on the table. Yes, every time the money went it, the short stack doubled, but it was more than that. Nobody was that out of line in their play.

When a bubble that long finally bursts, what might otherwise have been a decent structure has descended into a crapshoot. Regardless of what the internet kid might think, that’s not in his interest. If he’s genuinely better than the opposition, he’s barely going to be able to exploit that superiority if the average stack has dwindled from about 40BBs to 20BBs.
Let’s do some maths on this. The following covers small (100 runners or fewer) side events where:

a) Bubble factors are high – the smallest prize is often two to three buy-ins.
b) The bubble is at the final table where it’s easy to discuss bubble prizes with all the players left.
c) Unlike most main events, bubbles last longer because there aren’t dozens of min-cash prizes to get through before the big money (which would lower the bubble factor).

Imagine the players agreed to take one buy-in off first and second combined, as a bubble prize (in a 60%/40% split). The maximum downside for any player is 0.6 buy-ins. On a bubble with, say, seven prizes, let’s assume the chip leader has a quarter of the chips. What is he conceding with this move? Very approximately about a quarter of a buy-in (see footnote). What’s he gaining in exchange for this loss?

This is difficult to put any definite figures on, but conceptually he gains the opportunity to exploit and exert pressure on his opponents with a much bigger, more effective stack after the bubble bursts. Let’s assume that an “unburst” bubble lasts about one level longer on average in small side events than one where a bubble prize has been agreed. So, on average, the post-bubble effective stacks will be reduced by about a third when a bubble prize hasn’t been agreed. It’s tough to assess how much effect that will have on edge, but given that players will be gunning for a first prize of around 25 to 30 buy-ins (in a 70-runner tournament), it’s certainly conceivable that this is worth more than a quarter of a buy-in.

Of course, most players – even the decent ones – only know what they’ve read in the textbooks: that agreeing to bubble prizes is wrong.
The factor I’ve missed off my calculation is the gain, made by those refusing a bubble prize, of exploiting players who play poorly on the bubble. On balance (and once again, without any explicit calculation), I’d say this is still the biggest factor providing that there are players who are playing way off optimal bubble strategy. As I say, I’m not convinced there are so many of these players left out there. My advice is to play the bubble for about half a level, gauge who is doing what, and maybe agree to a bubble prize if it seems like no one is making any big mistakes.

Another option is for the tournament organisers to do something. Bubble factors are mainly a function of the differential between the prizes in the next pay jump. On the bubble of a satellite, this is set and difficult to manipulate. But what about side events when the bubble prize can often be over two buy-ins? Shouldn’t tournament organisers be experimenting with smaller pay jumps at the bottom end of the prize pool? Realistically, they don’t care either way – whatever they gain by speeding up the bubble will be lost by post-bubble play going on longer. However, it would be interesting to see whether players preferred smaller pay jumps at the lower end of the prize pool in this kind of tournament.

Even this is an approximation, but the chip leader has about a 25% chance of coming first, so he loses 0.25 x 0.6 buy-ins there; contingently, he also has around a (100%-25%) x 25% chance of coming second, where he loses 0.4 buy ins (so 0.75 x 0.25 x 0.4 buy ins); but don’t forget to add the chance that he bubbles (about 1%?), where he gains 1 buy-in. This rough guess works out at around 0.21 buy-ins lost.

Tags: Pickleman, strategy, bubble play