limp fold stat not ok

[kflug]

The Preflop Limp Fold stat by WhiteRider on page 3 of the repository has a couple faults. First it counts all the times a player limped and then folded Preflop, But included in that are times when the player may have 3bet or even 5bet before folding. quite different from limping and folding to a single raise.
Secondly the stat is expressed as a percentage of the times player limped rather that a percentage of the times player limped and faced a raise.
WhiteRider did correct the second error in a post about making limp fold stats but the repository stats are not updated.
The first fault may seem nitpicky but it is important i believe because you want your limp/fold, limp/call and limp/reraise pcts to add up to 100

[WhiteRider]

Thanks for pointing that out - I'll investigate and update as needed.

[WhiteRider]

First thought - we would need to check for limp/fold, that there is 1 call (the limp) and no raises.

sum( if[holdem_hand_player_statistics.flg_p_limp AND holdem_hand_player_statistics.cnt_p_call=1 AND holdem_hand_player_statistics.cnt_p_raise=0 AND holdem_hand_player_statistics.flg_p_fold, 1, 0] )

This raises an issue with limp/call and limp/raise.
I'm not sure whether it's possible at the moment to separate cases where you limp/call/raise from when you limp/raise/call.

We can, of course, find limp/call (with no raises) and limp/raise (with no calls), but when we have both raise and call involved we can't tell what order they happen in. (I believe the info is in the DB, but the required lookup function isn't yet available)

[kflug]

Here is a hand history:
Full Tilt Poker Game #55555558888: Table Melvin (deep hu) - $3/$6 - No Limit Hold'em - 17:41:33 ET - 2008/10/06
Seat 1: --DONKEY1965-- ($978.15)
Seat 2: ai2499 ($333.80)
Seat 3: tester ($999)
--DONKEY1965-- posts the small blind of $3
ai2499 posts the big blind of $6
The button is in seat #3
*** HOLE CARDS ***
tester calls $6
--DONKEY1965-- calls $3
ai2499 raises to $24
tester calls $18
--DONKEY1965-- folds
*** FLOP *** [6s Ah 8d]
ai2499 bets $20
tester folds
Uncalled bet of $20 returned to ai2499
ai2499 mucks
ai2499 wins the pot ($53.50)
*** SUMMARY ***
Total pot $54 | Rake $0.50
Seat 1: --DONKEY1965-- (small blind) folded preflop
Seat 2: ai2499 (big blind) collected ($53.50)
Seat 3: tester folded on the Flop


I can't see anything wrong with this history yet when i load it into a test db. 'tester's limp is missed completely the flg_p_limp is false and the limp call % is Zero. If I run the report from Donkey's point of view he scores a limp and a limp/fold.

This raises an issue with limp/call and limp/raise.
I'm not sure whether it's possible at the moment to separate cases where you limp/call/raise from when you limp/raise/call.

We can, of course, find limp/call (with no raises) and limp/raise (with no calls), but when we have both raise and call involved we can't tell what order they happen in. (I believe the info is in the DB, but the required lookup function isn't yet available)

For what it's worth i don't think the distinction is 2 important. I count any hand where a player limped and then later came out of the wood work and reraised. as a limp reraise. and any hand where the player limped and just called a raise at least once without ever reraising as limp call