Fold to CBet in the BB vs a CBet from BTN

[oracle3001]

I am wondering if / how one would possible to have a series of custom stats along the lines of

BTN cbets (lets say after RFI pre and BB calls*), want to know how often BB folds to this.

* the preflop sequence for the moment isn't crucial, rather getting a handle on how one would define the flop action i.e BTN cbets and BB (who is OOP) folds.

Now obviously there is the IIP vs OOP and position filters which I could set for BB, but this doesn't take into account the position of the cbettor. If I set position = BB, OOP facing cbet, for the BB that would include cbets from all other positions except the SB.

So I want to set up a series of custom stats that are along the lines of

SB fold to cbet vs cbet from BTN
BB .............. vs ........... BTN
SB .............. vs ............ CO
BB ............... vs ............ CO

basically how we can already do for 3bet preflop, where by one can form a matrix of 3bet position vs position

[kraada]

The way to do this would be similar to the preflop method - create a Fold to BTN Cbet stat and then change the Position property to BB.

To create your fold to btn cbet you'd start with fold to general cbet then specify that the first bet on the flop came from the button. The way to do this is to use the string cash_hand_player_statistics.str_aggressors_f - this string contains the list of all positions who made aggressive actions on the flop. So the first value is always the position of the person who made the bet. Thus if substring(cash_hand_player_statistics.str_aggressors_f from 1 for 1)::int = 0, the bet came from the button. And if we faced a cbet that cbet must have come from the button as well. (This aggressors string exists for all streets in PT4 and is awfully useful.)

[oracle3001]

Looking at the 3bet vs btn_2bet and 3bet vs co_2bet columns they use

cash_hand_summary.str_aggressors_p LIKE '80%', 1, 0])

cash_hand_summary.str_aggressors_p LIKE '81%', 1, 0])

obviously I understand the aggressors_p bit, but what is the LIKE '80%' and "LIKE '81%' mean? And could I use the _f version in my new columns?

e.g. something like this for the altered cnt_f_cbet_def_action_fold part of the stat

sum(if[cash_hand_player_statistics.enum_f_cbet_action='F', 1, 0] AND cash_hand_summary.str_aggressors_f LIKE '80%', 1, 0]))


I also want the opposite of this

i.e cbet flop % of BTN when BB is in the pot. etc

although I am thinking that CBet Flop (with position = BTN, position = IP set) will pretty much do the job, as that will be all situation vs the blinds, and the same will be true for position = CO.

[kraada]

The preflop string has one special characteristic - the first preflop aggressor is treated as the big blind - so every preflop string starts with an 8 (this has the benefit of making the length 2 for 2bet pots, 3 for 3bet pots, and the 3rd character the position of the 3bettor, and stuff like that).

% in the LIKE is a wildcard - it can be any or no characters. So '80%' gives us the 8 that it always starts with, the 0 means that the button made the first raise, and anything can happen afterwards. '81%' again starts with 8, but then 1 means the CO made the first raise, and again anything can happen afterwards.

You can do any kind of string matching (including this matching) for the flop (_f) version of the string (or turn or river which are _t and _r respectively).


sum(if[cash_hand_player_statistics.enum_f_cbet_action='F', 1, 0] AND cash_hand_summary.str_aggressors_f LIKE '80%', 1, 0]))

wouldn't really do what you want, you'd want to put everything before the ",1, 0]" piece like so:

sum(if[cash_hand_player_statistics.enum_f_cbet_action='F' AND cash_hand_summary.str_aggressors_f LIKE '80%', 1, 0]))

This would mean that the player in question folded to a flop cbet and the flop action was: bet from the BB, raise from the button (and anything could happen after that).