Tutorial: Creating C-bet Success Statistic

[BoFo]

Thanks, it now seems to work. U are not only kind, but also rather bright.

Yours, Bo.

[WhiteRider]

LOL
YW.

[BoFo]

There is one more thing: it comes out like 52.44 and two squares thereafter. Those cant be copied into this message, but I hope you know the type. It's like when I try to open something in "Word" that the program does not recognize. I have got /%.2f in both fields.

Yours, Bo.

[WhiteRider]

Yes, I understand what you mean.
Please Export your stat and attach it to a ticket on the Support system , and link to this thread in the reference field of the ticket so that we can investigate.
Let me know your ticket number.

[BoFo]

Support Ticket #37520

[WhiteRider]

I replied to your ticket - you have pressed <enter> at the end of /%.2f, that is what the boxes are trying to represent. If you remove the newline it should remove the boxes.

[BoFo]

Bravo! It does not take much to mess it up! Thanks again!

[WhiteRider]

No problem.

[feint06]

Very cool tutorial, should be stickied imo.

So what is our test? First we make sure we did c-bet (flg_f_cbet) and we didn't see a turn using NOT and (flg_t_saw) and we never faced a raise (NOT(flg_f_face_raise)).

Why this test? It should be obvious why we want to make sure we c-bet. But then what is a success? All opponents folded. Unfortunately there is no flg_all_opponents_gave_up flag. But what happens when all opponents fold? We don't get to see the turn. Then you have to make sure there are no cases you are missing. If we only used the first two, we would miss the situation where we bet, the other player raised, and we folded, thus not seeing a turn, but our c-bet would most definitely not have succeeded

Are there any other cases where we c-bet, don't see a turn and don't get raised that the c-bet fails? I can't think of any.

This is true, but I think your definition also undervalues the cbet success a little. There will be times when we cbet, an opponent raises, and we 4bet causing the opponent to fold (example -- you have AA and make a standard cbet on a raggy flop. A competent villain knows you could be cbetting with ace-high and takes a stab at the pot by 3betting -- you shove and he folds). These cases will be counted as failed cbets (which I suppose they are since the cbet itself failed) but the end result was a profitable situation (which I believe is the goal of this statistic).

So it would seem to me that to include these scenarios, we need a definition that is based on us winning the pot on the flop instead of us not seeing a turn with some other stipulations.

[kraada]

True, in that case use holdem_hand_player_statistics.flg_won though make sure to have not(turn saw) otherwise you end up getting all hands you won that you cbet and that's going to be such a lot of hands that it's going to be hard to gauge how well the c-bet worked.

Personally I'd rather make a bet/3bet statistic and see how often that wins; the idea behind this stat was to see how often your cbets just win without further action, to answer the question of how often you should cbet regardless of cards/board texture/etc. If you want a stat for slightly different reasons that changes how you'd want the definition.