Calculation of CBF / CBC / CBR

[GrumpyB]

I have just added the fold to continuation bet (CBF) / call c-bet (CBC) / raise c-bet (CBF) stats to my HUD.

I expected these to add up to 100%:

50/50/0
75/0/25
67/-/33
75/25/0

but then I see some people with really strange figures, like this real example:

72/36/62 after 180 hands

Now I am guessing that this could happen when a player raises a c-bet, gets re-raised and then calls or folds to a re-re-raise (which may not be so uncommon if he raises 62% of all c-bet!) with both the raise and call/fold counted against one c-bet. But just wanted to check that this guess was right!

thanks

[APerfect10]

Your understand is absolutely correct and quite impressive :) Most people think that the stat should add up to 100, which may appear logical, but is not the case. Nice work!

Best regards,

Derek

[GrumpyB]

Thank you sir - I was worried it was going to be a really dumb question.

And when I said "I have just added the fold to continuation bet (CBF) / call c-bet (CBC) / raise c-bet (CBF) stats to my HUD" - what I really meant was "I have just stolen APerfect10's PAHUD layouts from a threa I found and it contains ....." so I am doubly happy that it's you that answer.

Pinching your hud has also switched me from AF to Aggression Frequency - and I think I like it :D

[APerfect10]

Good deal :)

Aggression Frequency is a much better stat to use in your head-up display. I think you will find it more useful.

The only time I think that Aggression Factor is advantageous is when doing statistical analysis on groups of players...

Best regards,

Derek

[WhiteRider]

I can see why that wouldn't have to add up to 100%, but should the response to a 3-bet after a c-bet really count as responding to a c-bet?
I'm not sure I want those stats adding together!

As an extreme example; if someone always raises a c-bet but then folds if re-raised, their 'fold to c-bet' stat would be quite high (up to 50%), when actually they NEVER fold to a c-bet only to a c-3-bet. Likewise, their 'raise c-bet' stat would be only 50%...

[APerfect10]

No, if they raised every cbet then folded to the 3bet then their 'Raise CBet' would be 100 and their 'Fold CBet' would also be 100. This is precisely why the stats do not add up to 100 :) In your scenario, the stats would always add up to 100 and you proved exactly why that would be inaccurate.

Best regards,

Derek

[WhiteRider]

OK, forgive me - it's been a long day and I knew I shouldn't have posted till tomorrow, but now I've started...

In one hand of my above example, are they facing one c-bet or two? Or is it just counted as a hand where they were c-bet into? (I think it's the second from the above Q&A)

Regardless of me misunderstanding that it would report 100% instead of 50% - I still hold with my statement that I don't think the fold to a c-3-bet should count as a fold to c-bet! That's very misleading!

[APerfect10]

It's one cbet and I understand your problem with the stat. Either it counts towards the cbet defense or not. While it may be misleading to include it, I think it would be even more misleading to ignore it so we are stuck doing one or the other.

Best regards,

Derek

[WhiteRider]

This should probably be in the PT3 forum, but I'll continue here since we've started..
I was thinking about how this would work in PT3, and hoping that we will be able to extract the 'pure' facing-cbet stat, but it looks like PT3 already does it as I'd originally expected. Can you please clarify the following logic?
(All in "holdem_hand_player_statistics")

PT3 DB has columns [bp1p]flg_f_cbet_def_opp[/bp1p] (boolean: faced continuation bet) and [bp1p]enum_f_cbet_action[/bp1p] (character: action when facing cbet).
Since enum_f_cbet_action is a single character, I assume that this only gives one possible response to a cbet per street?
(so a c-bet response stat calculated from this would total 100%)

There are also flg_f_3bet_def_opp and enum_f_3bet_action.
In order to get the same cbet response stat we discussed above you'd have to combine these two, so maybe it won't be being done the same by default? Either way, I should be able to access the "pure" cbet response stat that I want?


Further note: I see flg_f_bet_def_opp but not enum_f_bet_action.
Ah, sorry - I think that is a typo in the docs - I guess it should be flg_f_3bet_def_opp as I quoted above?

There is a column flg_f_face_raise. Raise is a 2-bet, yes?
This then suggests the question, should there be a flg_f_face_bet (or, to keep the previous naming convention, flg_f_bet_def_opp and flg_f_raise_def_opp) and associated enum_..._action stats?

If there are to be stats like this, maybe this is a way of including individual betting amounts - e.g. amt_f_bet, amt_f_raise, amt_f_3bet, amt_f_4bet?

[APerfect10]

[quote="WhiteRider"pso]This should probably be in the PT3 forum, but I'll continue here since we've started..
I was thinking about how this would work in PT3, and hoping that we will be able to extract the 'pure' facing-cbet stat, but it looks like PT3 already does it as I'd originally expected. Can you please clarify the following logic?
(All in "holdem_hand_player_statistics")

PT3 DB has columns [bpso]flg_f_cbet_def_opp[/bpso] (boolean: faced continuation bet) and [bpso]enum_f_cbet_action[/bpso] (character: action when facing cbet).
Since enum_f_cbet_action is a single character, I assume that this only gives one possible response to a cbet per street?
(so a c-bet response stat calculated from this would total 100%)

There are also flg_f_3bet_def_opp and enum_f_3bet_action.
In order to get the same cbet response stat we discussed above you'd have to combine these two, so maybe it won't be being done the same by default? Either way, I should be able to access the "pure" cbet response stat that I want?[/quotepso]

That sounds about right to me...

[quotepso]Further note: I see flg_f_bet_def_opp but not enum_f_bet_action.
Ah, sorry - I think that is a typo in the docs - I guess it should be flg_f_3bet_def_opp as I quoted above?[/quotepso]

I will look into this and get it fixed.

[quotepso]There is a column flg_f_face_raise. Raise is a 2-bet, yes?
This then suggests the question, should there be a flg_f_face_bet (or, to keep the previous naming convention, flg_f_bet_def_opp and flg_f_raise_def_opp) and associated enum_..._action stats?[/quotepso]

Yes, flg_f_face_raise is a 2bet. I do not think that is needed since the action facing a normal bet would be either call, fold or raise which is already noted and assumed you are facing a normal bet. Otherwise, if it is a raise, 3bet, etc, there are the appropriate flags to distinguish this. This is off the top of my head, I may need to re-look into it to confirm.

[quotepso]If there are to be stats like this, maybe this is a way of including individual betting amounts - e.g. amt_f_bet, amt_f_raise, amt_f_3bet, amt_f_4bet?[/quotepso]

Check out holdem_hand_player_detail

-Derek