♠ Saturday, October 16, 2004
Thoughts on PartyPoker's Bad Beat Tables
I can't be the first person to think of this, but I don't know if any of the blogs have covered it. I'm not caught up on the blogs, not by a long shot.
In any case, my point is this. It seems that the bad-beat tables on PartyPoker are ripe for setting up a consortium to win it. With the rules set as they are at the moment, where 8888 or better must be beat, a full table of people whose only goal is to win the BBJ would fold everything other than 88 or better, or suited connectors to three-gappers. Nothing else can make a qualifying hand. (Actually, someone playing A8 who hits an 88899 board and loses to 99 might be eligible for the jackpot. Cardrooms differ on this, and I'm unsure how PartyPoker sees it.)
In any case, it's assumed that to get the most hands in as possible, and keep the rake down, that the players would check down the hand unless the board makes it clear that a jackpot is possible, in which case the players would need to bet it up enough ($20) that the rake is taken for the jackpot, which makes the hand eligible to win. If it folds around to the small blind, as many hands will, the small blind should raise a jackpot-possible hand and fold otherwise, to signal the big blind whether to continue with the hand.
A team playing in such a way would win more than their share of jackpots, at a fairly low cost to the players. At the lower limits, most hands would not be raked, and the pots won or lost would be tiny. On the other hand, "more than their share" of jackpots wouldn't mean "a lot of jackpots." Because of the more hands per hour, and because no jackpot-possible hands would be folded preflop, the likelihood that a consortium table would hit the BBJ might increase to between double and triple the likelihood that a "normal" table would hit it.
A consortium like this would probably have to consist of a good fifty people, to ensure at least one full table, and more when they can manage it. The payout when they do hit the jackpot would have to be split among them by some formula involving the total number of hours played by each member of the consortium, and the consortium's leaders would probably skim some administrative costs.
Now, here's the question: At what point is the BBJ high enough that all this is +EV? Always? Over $100,000? $200,000? $300,000? My gut says that if all went perfectly, then it's +EV at any level of the BBJ, but I haven't done extensive math on this.
The biggest snag, though, is that the actual winners of the BBJ would have to actually transfer the money to the head of the consortium. With the "big part" of the jackpot reaching over $100,000 on occasion, the temptation would be strong to tell the consortium to go hang. And so, unfortunately, the consortium probably wouldn't last beyond the first time it hit the jackpot.
When state lotteries' million-dollar games were reasonably new, meaning the 1980s, there were occasional reports of consortia attempting to buy up all the tickets to the larger drawings, guaranteeing themselves a win (or at least a split). You don't need a math degree to see that the pot odds would have to be three or four times the true odds to make this profitable, but that does happen sometimes, and you don't hear about consortia trying to buy up the tickets anymore. Is the whole concept just too unwieldy?
This would also be an ideal application (if you did the math on the EV) for a bot, cloned to play most if not all seats at the table, as it would only require a minimal amount of AI to make "correct" decisions.
This whole thing is probably academic, anyway, as if Party has any sense at all they have built into their terms and conditions the discretion to refuse to award the BBJ if there is suspicion of collusion.
As an aside, just read through your blog. Good stuff.
Other then that, love your blog.. keep up the good work...
And other times, I'm playing poker and staring at that $200,000 number scrolling ever-higher on Party's lobby screen, and I'm wondering ...
I've thought of a bot, too; I used to be clever enough to write such a program, back in the Commodore 64 days, but now it would take a loooot of work to re-immerse myself in whatever the language is these days, and moreso to figure out how to work a Windows machine with that language. But it's probably not this:
2120 FOR I=0 to 255:POKE 49642, I:NEXT
Anyway, yes, bots good in this case.
But here's another question: Would PP welcome attempts to hit the BBJ with a consortium, or not? You wouldn't be depriving anyone of their chance to win, just increasing your own chances. And, you'd still be contributing to the rake and BBJ, although with perfect play you wouldn't be contributing much. They don't seem to mind when players say in chat "everybody fold everything worse than 88," but would they truly mind a consortium?
"The BEST hand of the winner and the loser must include both the hole cards and the hand must go to a showdown."