Contributed Rake

[Wolfskin]

Is it possible to run this query in pokertracker?
It's contributed rake and not weighted contributed like posted earlier in this thread.

Code: Select all

SELECT hs.amt_rake / COUNT(pd2.amt_bet_p) AS contributed_rake FROM holdem_hand_player_detail pd
JOIN player p ON p.id_player = pd.id_player
JOIN holdem_hand_player_detail pd2 ON pd.id_hand = pd2.id_hand
JOIN player p2 ON p2.id_player = pd2.id_player
JOIN holdem_hand_summary hs ON pd2.id_hand = hs.id_hand
JOIN lookup_sites ls ON ls.id_site = hs.id_site
WHERE p.player_name like '%XXX%'
AND ls.site_name like '%XXX%'
AND pd.amt_bet_p >0
AND pd2.amt_bet_p >0
GROUP BY pd.id_hand, hs.amt_rake, ls.site_name;

[WhiteRider]

You can't run queries directly in PT3, but you can use PGAdmin: Start > Programs > Postgres > PGAdmin.
Select your database and click the SQL icon in the toolbar.

[Wolfskin]

Is it possible to make a custom stat out of this query? I was only able to do it in SQL and don't know how to do it in pokertracker 3.

[kraada]

Which stat section did you want to have this stat in?

[Wolfskin]

It would have to be an individual hand statistic, because it isn't a sum. When I tried to SUM it with a COUNT, SQL gave an error, that aggregate functions won't work.

Are "JOIN" commands possible in pokertracker?

[kraada]

You can run individual PostgreSQL queries in the custom reports section (see this post), but not in the custom statistics section.

There is also the option to put it in Holdem Cash Player Statistics so it sums over all hands and isn't for each individual hand; I wasn't sure what you wanted to do with the stat and the way you'd built it for each will vary slightly.

[Wolfskin]

This would be better, because it gives an integer instead of a real. It's the number of contributing players.

Code: Select all

SELECT COUNT(pd2.amt_bet_p) AS contributing_number_of_players FROM holdem_hand_player_detail pd
JOIN player p ON p.id_player = pd.id_player
JOIN holdem_hand_player_detail pd2 ON pd.id_hand = pd2.id_hand
JOIN player p2 ON p2.id_player = pd2.id_player
JOIN holdem_hand_summary hs ON pd2.id_hand = hs.id_hand
WHERE p.player_name like '%XXX%'
AND pd.amt_bet_p >0
AND pd2.amt_bet_p >0
GROUP BY pd.id_hand;

[kraada]

I'm a bit confused about what you want now so let's try to clarify.

Which do you want?

(a) A stat for individual hands that shows for this hand, the player contributed to the rake.
(b) A stat that shows for an individual player, over all of his hands, how much rake total there has been in pots he contributed to.

I'm not sure what the number of contributing players query would be useful for here, so if there's a (c) please elaborate.

[Wolfskin]

I'm a bit confused about what you want now so let's try to clarify.

Which do you want?

(a) A stat for individual hands that shows for this hand, the player contributed to the rake.

This can already be checked, by using holdem_hand_player_detail.amt_bet_p > 0

(b) A stat that shows for an individual player, over all of his hands, how much rake total there has been in pots he contributed to.

This can already be done by using sum(if[holdem_hand_player_detail.amt_bet_p >0, holdem_hand_summary.amt_rake, 0])

I'm not sure what the number of contributing players query would be useful for here, so if there's a (c) please elaborate.

to calculate the contributed rake, which depends on the number of players, that have money in the pot, which is for an individual hand holdem_hand_summary.amt_rake/number_of_contributing_players
when you have the number of contributing players you could calculate the contributed rake, but I guess you would need a new stat, because it requires a join to get the information
but it isn't that important, when it can't be done right now in pokertracker

[kraada]

You can't be perfect about it, as someone could limp/fold but if you use holdem_hand_summary.cnt_players_f that will be all players who reached the flop, which is a pretty good approximation. Also, and it may no longer be this way, but certain sites used to count you as contributing so long as you saw the flop, so that may actually end up being what you want (I know OnGame used to do it that way several years back.)