Know Thy Enemy - The PokerRoom FreeRoll Rankings - Last updated 2005-05-03
PokerRoom.com is a great place
to play poker online. It has many FreeRoll tournaments and other goodies that
let you play a lot of poker for little or no money. And, if you're a Macintosh
user, it's one of the few places on the net you can play online using your
Mac, because it has a java client.
Like many dedicated PokerRoom.com
FreeRoll moochers, I've often wondered who the best players really are. And because
I'm an old-fashioned hacker, it occurred to me that it would be a fun project to write
a program to grab tournament data from PokerRoom.com
and analyse it.
Before I get into details, please note that this is a very large page and
you must have javascript enabled to see it (since I use javascript to compress
the results tables to a reasonable size). Hopefully it will be finished
loading by the time you've read the preface!
Okay, so what I did was write a program in RealBasic
that logs into my PokerRoom.com account
and requests tournament histories. It uses the weekly calendar pages as starting points to find
the tournaments. While I can always get tournament histories for tournaments listed on recent
weekly calendars, not all of the scheduled tournaments are listed (for example, the 4:15 FreeRolls).
In order to grab them, the program sniffs around at adjacent tournament numbers looking for other
tournaments that it otherwise might have missed, and so it can find most -- but not all -- of the
scheduled tournaments.
Since I've just started grabbing the tournament histories, and since tournament histories
seem to "erode" after a few weeks (try scanning back through the weekly calendars and you'll see
what I mean), the further back in time I go, the spottier my coverage. I did manage to grab some
tournaments as far back as December 2004, but alas, not the FreeRoll that I won. Oh well! As I
continue to add tournaments in the weeks to come, the rankings should slowly become more
comprehensive. Note that this page only lists FreeRoll rankings; if there is sufficient interest
I can put up a page for all scheduled tournaments, but it's going to be huge!
At this point, I think I should mention to anyone considering doing a similar scanning
project that it is important to be polite to your data provider. For example, my program
only made one request at a time to the PokerRoom.com server, and I set it to back off when
it ran into a block of Sit-N-Go tournaments. While it would have been very interesting to
generate rankings for the SNG players, there's such a huge number of SNGs running every day
that I wouldn't feel comfortable trying to grab the data without explicit approval from the
PokerRoom.com staff.
Once I had the data, I had figure out how to compute rankings of the players. I came up
with two ways to do it; the first is simply "total amount won", while the second is "Return on
Investment" -- how many dollars you win for each dollar invested.
Now some of you may be saying, "How can you have a return on investment when you don't
pay anything to play in the FreeRolls?". Good question! What I did was assume that the "Buy In"
for a FreeRoll was the total of the prizes divided by the number of entrants. So for example,
if there were 1000 entrants in a $1000 FreeRoll, it could be treated the same as a $1+$0 tournament.
A similar technique can be used for tournaments that either require a ticket to enter, or
cost money to enter but give tickets (and possibly cash) as prizes. In each case, you can compute
the implied value of the ticket in order to determine a return on investment. Note that you can't
easily do this with tournaments that both require a ticket and pay out a ticket, but these are
very few in number, so I just ignored them.
So here are the FreeRoll rankings. The Return on Investment rankings only list players who
participated in 10 or more FreeRolls, because ROI requires a bit of a track record. #-T is the number
of tournaments entered.
