Nat Arem's Blog

So in case I haven’t made myself clear in past blog entries, I am going to attend Emory Law School this fall. I think orientation starts around the 25th of August and I plan on getting down there around the 21st. I’m very excited to move to Atlanta and get started with this next chapter in my life. I still plan on being involved in the online poker world, but I’m going to have to pull back in a number of ways just due to the time requirements of a law student. I won’t be able to read P5s all day, mess around in 2+2 OOT or spend all day on AIM. It’s unfortunate, but it’s what I have to do if I want to be successful in law school. I still plan on running thepokerdb and I’m working on a deal that would not only make the db much more detailed, but would also lower my overall time commitment.

I am going out to Vegas for the WSOP one week from today on the 24th and I’ll be staying at the Bellagio with my good friend Ozzy Sheikh, aka Ozzy 87. I plan on spending a lot of time hanging out with people from the various “houses” around Vegas. I think schaefer, colson10, ackbleh, Bushman, Martine23 and a few others are in one house. Genius28, BeL0WaB0Ve, Empire2000, Green Plastic, ahh_snap and others are in another house. I’m sure there are more as well. I plan on getting some fun home games going, hopefully with some crazy games. After Vegas and a short family trip, I’m going to be holed up studying for the rest of the fall (and the bulk of the next three years), so I plan on having fun. I also plan on being at the P5s party (obv), so please introduce yourself if you see me there.

… of like low to middle limit ring poker.

Here are the requirements:

1) Must play at least 5/10 limit or 1/2 no limit.
2) Must play at least 20,000 hands a month, preferably a lot more. Ability to play short-handed is also preferred but it’s not required.
3) Must be comfortably bankrolled and not on the verge of going broke.

That’s about it. Let me know if you or someone you know might fit these requirements. Also, I have to either know you or you have to know someone who I know who can vouch for your claims. I only need one person for this particular project.

I can be contacted at natarem@gmail.com. Thanks.

EDIT: I think I found a player for this project, I will post again if something falls through.

Argh, I just misclicked and lost my full version of this blog entry. Long story short, this site Pokerbility wants to advertise with thepokerdb. I noticed something funny on the site, despite it looking very professional.

This image, which is featured prominently:

Anyone else think it’s odd that trips (Q6o on a 656 board) is listed as a full house? Or that they call it a Sklansky Group 1 hand (which I hope they got permission to use, otherwise they might have some shit coming to them)?

Oddly enough, I think I’ll pass on this particular piece of software to be represented on the site. The sales guy seems like a nice guy and I’m sure it’s not his fault, but this is kinda disgraceful.

Anyway, I wrote the following response to his initial email:

Why does it say I have a full house? http://www.pokerbility.com/img/control.jpg … and why is it group 1?

Nat

His response:

Hi Nat,

I’ve forwarded your email to our support team. I will receive an answer from them soon and send it to you. This problem for me seems to be really strange. We had no such problems before but if the problem is calculator it will solved in the shortest possible time.

I’m sure they’ll replace the image soon, but it’s funny while it’s up there. I’ve got it saved just in case. This doesn’t bode well for their software.

EDIT:

Just got this reply.

Hi Nat again,

I’ve got a reply from support team now. I didn’t pay attention that you took this picture from our website. This is just a picture the graphical designer has created for the application main screen. It is not a screenshot of the real application. Therefore the statistics are wrong and the group is shown. We will take care of it and will change the picture on the site. There is nothing wrong with the software statistics.

I can appreciate that and all, but what if an online poker site had a promo graphic with two of the same card on board? Or if the pot was being shipped to the wrong hand? Shouldn’t these types of things be taken care of to be graphically accurate?

ANOTHER EDIT: They fixed the picture! Actually, this thing looks kinda useful especially for a novice. Maybe I will consider putting up an ad on thepokerdb.

If you want to know the hilarious story behind the above pic, read this thread. I crack up every time I read it.

Click on the thumbnail to see the full lobby.

This will probably confuse a lot of my readers, but I’ve been looking at this a lot lately…

The PokerTracker Data Schema

You can also get a good overview by looking at this image (it’s big).

Basically, PokerTracker has to take a bunch of text and organize it into variables, which can then be queried to look at compiled data. That’s the parsing process. But then it needs an organized way to store those values and that’s where this schema comes in. I count 25 tables that it uses to store the data, which is about the same number that we use over at thepokerdb. But we have duplicate tables for each site, a user table, a session table, etc. In terms of storing actual tournament results, we only need a few tables. Storing hand histories results is a lot more complicated.

If you actually go into PokerTracker and double-click on any column header, it will take you to the “SQL Debugger” where you can view the query being used for that column.

Here is the query under the “General Info” tab double-clicking on the hand listing columns for player H@££INGGOL in Stars $10/$20 NL:

SELECT hole_cards as starting_hand,
game_players.card_order1 as card_order1,
game_players.card_order2 as card_order2,
game_players.card_order3 as card_order3,
sum(game_players.won_hand) as won_hand,
sum( CASE WHEN game_players.won_hand = 1 THEN game_players.saw_flop_n ELSE 0 END ) as won_wsf,
sum((total_won – total_bet)) as amt_won,
sum(game_players.big_blind_n) as dealt_bb,
sum(game_players.small_blind_n) as dealt_sb,
sum(game_players.saw_flop_n) as saw_flop ,
sum(game_players.pre_flop_raise_n) as raised_pf,
sum(game_players.raised_first_pf) as first_to_raise,
game_players.pair_hand as pair_hand,
game_players.connector_hand as connector_hand,
sum(game_players.vol_put_money_in_pot) as vol_saw_flop,
sum(game_players.limp_with_prev_callers) as limped_in,
sum(game_players.went_to_showdown_n) as went_to_sd,
sum( CASE WHEN game_players.went_to_showdown_n = 1 THEN game_players.won_hand ELSE 0 END ) as won_sd ,
count(*) as times_dealt,
sum(cold_call_pf) as times_cc,
sum((total_won – total_bet) / (game_level_big_bet + (game_level_big_bet * pl_nl))) as bb_won
FROM game, game_players, game_level
WHERE ( game.game_id = game_players.game_id ) and
( game_players.player_id = 1048 ) AND
( game_players.hole_card_1 is not NULL ) AND
( game.game_level_id = game_level.game_level_id )
AND ( game.game_level_id IN (6) ) AND ( game.site_id IN (2) ) GROUP BY game_players.hole_cards, game_players.card_order1, game_players.card_order2,
game_players.card_order3, game_players.pair_hand, game_players.connector_hand

Some basic SQL commands are used here, such as SELECT, WHERE, AND, GROUP BY, SUM, etc. (it’s conventional to put the SQL in caps and variables in lower case, but that’s not always the case)

A really basic SQL line might look like this: SELECT player_id FROM table_id WHERE name = ‘N 82 50 24′

That’s basically saying get the “player id” associated with the name “N 82 50 24″ from the “table_id” table. Obviously the PokerTracker SQL is a lot more detailed and is SELECTing a lot more information, but the basic concept is the same.

Anyway, I find this stuff interesting and I want to explore ways to make hand history data schemas more advanced and capable of more detailed information extraction from parsed hand histories. This schema might be as good as it gets and capable of just about anything, but I want to find out if that’s the case.