4 bet , then fold

[kflug]

I'm working on some stats for HU play and I'm stumped on the percentage a player 4 bets, then folds to a raise. I can determine how many times he did that. but I can't figure out how to get the denominator. ie... how often he 4 bet then faced a raise. Anybody got a suggestion how to find this amount?
Thx in advance.

[WhiteRider]

So I guess you know how to tell when you made a 4-bet (holdem_hand_player_statistics.flg_p_4bet), and you can also check for
holdem_hand_player_statistics.flg_p_4bet_def_opp (which is true if you had an opportunity to defend against a 4bet or higher - i.e. when you faced a 4-bet or higher).

[kflug]

sb-raise(2b), bb-reraise(3b), sb-4b, bb-allin, sb- call. A not unusual sequence in HU play. From the bb's perspective he/she faced a 4bet in the hand and made a 4bet in the hand. But we don't want this hand included in a 4 bet and fold stat because the bb didn't face a raise AFTER 4-betting and so didn't have opportunity to fold after 4-betting (5 betting in this case). Including this type of hand in the denominator of a 4-bet/fold stat would make it seem as though the 4bet/fold frequency was lower than it truly is. I suspect it would be much lower since these hands are fairly common. Any ideas?

[WhiteRider]

Sorry, yes - I didn't think about that case.
I'm not sure this is going to be possible - I can't think of a way of separating the cases at the moment.

If it is purely for heads up hands, then you may be able to do something by counting the preflop raises, using:
holdem_hand_player_statistics.cnt_p_raise
..although I can't actually see how it would help in this case! :oops:

I'll mention this thread to the development team and see if they have any ideas.

[kflug]

If it is purely for heads up hands, then you may be able to do something by counting the preflop raises, using:
holdem_hand_player_statistics.cnt_p_raise

I'm wondering about using the amt_p_raise_facing column to check if you faced a raise after 4betting. It seems to be the total amount of the last raise you faced in that betting round.. So perhaps sum(if [hhps_flg_p_4bet AND (hhps_amt_p_raise_facing > hhps_amt_p_raise_made OR hhps_cnt_p_raise > 2)], 1,0 ) would be the best effort at getting a count of hands where player 4 bet and faced a raise? Can anyone improve on that? Or think of a sequence that this stat would miss?
The second raise a player makes in 1 round is always at least a 4bet. I can't see any reason at first look why this stat might not also work for ring games.

[WhiteRider]

That looks feasible..
The best way I find to test this is to make your stat as you define, but then make a similar stat in the 'holdem cash hand' section too, and just remove the sum(if[ ... ]) parts so that it is just a boolean, and then format it as a checkbox stat which you can add to reports too.
e.g. I use the Hands tab and add the HCPS stat to the top report and the HCH stat to the bottom section.

See the 'clever formatting' part of the Custom Statistics and Reports FAQ for some cleverness with checkbox stats like this.

[kflug]

. I can't see any reason at first look why this stat might not also work for ring games.

It won't work for ring games because the players first raise might be a 4 bet.

[SukitTrebek]

i tried to make this stat myself last night. clearly it seems we need 4bet and 5bet flags separated.

my best attempt at hacking something together was to try to see when people 4bet and did not see flop. this seems like it should work and would only have the inconsequential effect of including times they 5bet and did not see flop. it wouldn't even need to be measured against the times they 4bet as this could be problematic (since it includes times they could 5bet), but simply measured per number of hands then multiplied by 1000 or something would still be helpful i'd think.

however whatever i'm doing with sql doesn't work. it's probably something minor.

i made a column for times they 4bet and did not see flop.

sum(if[holdem_hand_player_statistics.flg_p_4bet AND NOT holdem_hand_player_statistics.flg_f_saw , 1, 0])

it seems this is working, if i run a report just counting this, the people who 4bet more have higher values and vice versa.

but if i simply then take this column and divide by the column cnt_hands to create a useful stat, i get an invalid expression.

[WhiteRider]

sum(if[holdem_hand_player_statistics.flg_p_4bet AND NOT holdem_hand_player_statistics.flg_f_saw , 1, 0])

it seems this is working, if i run a report just counting this, the people who 4bet more have higher values and vice versa.

but if i simply then take this column and divide by the column cnt_hands to create a useful stat, i get an invalid expression.

So are you creating a Stat dividing your new column by cnt_hands?
Like:
(cnt_p_4bet_no_flop / cnt_hands) * 100

[SukitTrebek]

yes, it seems like it should be straightforward, but when i create a stat with the expression you have above, i get invalid expression.