(This is just a rough post that I’m writing up because of a quick aside in my previous post that wwants asked about it on twitter. You probably don’t care about this. But I don’t wanna ever be accused of missing an opportunity to nit pick, and I’m completely incapable of saying anything in 140 characters–or, frankly, 1400–so I wrote up a post.)
PTR had a pretty cool idea to compare the rake at various sites. I think that that’s wonderful. The sites have slightly different ways of charging rake which makes comparison hard and obscures what are almost certainly legitimate differences. Different sites divide the stakes and the number of players at the table differently, take rake at different increments, have different caps, etc. Some sites have BBJs, and Full Tilt even rakes run it twice pots differently to further confuse things.
However, this is actually a slightly deceptive problem because the obvious solution doesn’t really provide legitimate information. PTR just took the total rake generated in their database and divided by the number of hands (then multiplied by 100 to make the numbers look pretty). The really big problem with this is that all the sites charge rake only if you see a flop (except french sites…), and the amount depends on the size of the pot. So sites with more flops and larger pots–the type of site that almost everyone wants to play on–will tend to show higher rake by this metric, even if their rake structure is exactly equivalent.
This isn’t a trivial problem. For example, Cake, Stars, Bodog and FTP have identical rake at NL2000 FR except that FTP charges $1 extra for run it twice ($0.05 per $1 capped at $3) . But in PTR’s table, Bodog charges over twice as much rake as Cake, and there’s pretty good variation in general (Cake = $11.51, FTP = $13.80, Stars = $17.96 and Bodog = $23.35). Since almost every pot that sees a flop at 10/20 is going to be capped at $3 rake, this almost certainly means that players see the flop about twice as often on Bodog as on Cake (about 70% and 35% respectively). I don’t think that many players would consider that as a negative, and it’s certainly not Bodog’s fault. I’m sure that there are many similar examples.
A simple solution might be to just calculate the rake paid per flops seen. This seems roughly in line with how people think about the rake in the first place (“No flop, no drop”), and would likely give more accurate numbers. However, this still doesn’t deal with the problem that larger pots get raked more, which is likely a large effect at lower stakes where the cap isn’t reached as often. The best solution is unfortunately a bit more complicated: Take every hand played on all sites (or some representative subset) in a given game type and see how much rake each site would’ve charged for the same set of hands. RIT hands should be treated as though they have no extra rake, since players choose to run it twice. That’s not totally perfect, but the flaws are subtle and not very important (and you guys already think I’m a huge nit) so I won’t even bother to mention them. Extra points if PTR lets people see how much rake they would have been charged to play the same hands on a different site.
Incidentally, they also handle the BBJ incorrectly. Both Party and Cereus use the BBJ as an excuse to take extra rake from the players, each taking 10% of the jackpot every time it gets hit. Because of the way that they roll over a portion of the jackpot, they actually end up taking more than that–In the long run, Cereus takes about 13.3% and Party takes about 12.5%. That certainly costs players a significant amount of money, but PTR treats the BBJ as if players receive none of it back. This seems silly–Poker players are gamblers who are certainly willing to pay money for the chance of winning some down the road, and the popularity of BBJ tables makes it clear that players don’t think of the BBJ as increased rake with no benefits. This is quite unfair to Cereus and Party, and multiplying the BBJ rake by 13.3% and 12.5% respectively would be much better.
1 Comments.