Question Regarding 4bet Stat

[Acryl]

Hello! assuming someone opens 30% from a position and 4-bets 10% from that position. Does that mean he 4 bets 10% of the time when he gets 3 bet, so in other words 0.3*0.1 = 3% hand range or does it mean he 4bets a 10% hand range?

[Flag_Hippo]

The '4bet Preflop' stat is how often a player 4bets so this means they are 4betting 10% of the time when they have the opportunity to do so and this would also include cold 4bets. If you want this stat to not include cold 4bets then there is a '4Bet+ Preflop After Raising' built-in statistic (which is for 4bets or higher) and a '4Bet Preflop After Raising' custom statistic available in the Download Warehouse (which is specific for 4bets). The 4bet range after raising would be their 2bet% x 4bet% and there is a '4Bet+ Ratio' built in stat available in PokerTracker 4 which is similar and a '4bet range (RFI * 4Bet)' custom stat in the Download Warehouse.

[Acryl]

The 4bet range after raising would be their 2bet% x 4bet%

Thanks for your answere. Just to make sure I get it right: if their 2bet% range was 30% and their 4bet% was 20%, it would mean that out of a 30% opening range, he would 4bet 6% of hands, right? So basically, 2bet and 3bet% are always directly their exact range but once its a 4bet or higher we have to convert it or use a different stat (one of those you provided), correct?

[Flag_Hippo]

So basically, 2bet and 3bet% are always directly their exact range but once its a 4bet or higher we have to convert it or use a different stat (one of those you provided), correct?

Yes that's basically right but this wouldn't apply to cold 4bets e.g. if a player is cold 4 betting when there has been a 2bet & 3bet in front of them then their cold 4bet% is equal to their cold 4betting range%.

[Acryl]

okay thanks, got it!

[leviathan74]

The 4bet range after raising would be their 2bet% x 4bet% and there is a '4Bet+ Ratio' built in stat available in PokerTracker 4 which is similar and a '4bet range (RFI * 4Bet)' custom stat in the Download Warehouse.

Sorry to bump an old thread, but for a long time, I 've been scratching my head trying to understand what stat gives the most accurate number for the range of hands an opponent 4bets and google always brings me to this thread.

Right now I am looking at those stats and they give me slightly different numbers; basically, there's a 2% discrepancy in those numbers with 4bet ratio giving me a wider range of hands that people 4bet with than the rfi*4bet stat. Since 4bet ranges are so narrow, it makes a huge difference if people are 4betting 4% of hands rather than 2% of hands!

I looked at the stat definitions and 4bet ratio is using PFR% as the numerator and 4bet opportunities as the denominator while the latter is using RFI% as the numerator and 3bet defense opportunities as the denominator.

Assuming that people are 4betting the same range of hands regardless of whether they 4bet as the PFR and as the RFI (which isn't necessarily the safest assumption), what explains the discrepancy? More to the point, in what way 4bet opportunities produce different results than 3bet defense opportunities? Is it because you re facing all ins and you can't 4bet which artificially makes the 4bet range seem more narrow?

Fwiw, I went to the rfi*4bet stat and replaced 3bet defense opportunities with 4bet opportunities and I got even wider range numbers.

I am perplexed. Any ideas?

[Flag_Hippo]

Right now I am looking at those stats and they give me slightly different numbers; basically, there's a 2% discrepancy in those numbers with 4bet ratio giving me a wider range of hands that people 4bet with than the rfi*4bet stat.

Please note that this statistic in PokerTracker 4 is called '4Bet+ Ratio' and it is not specific to 4Bet hands only. The custom statistic called '4bet range (RFI * 4Bet)' in the Download Warehouse is also not specific to 4Bet hands only - it is 4Bet or higher. If you want a statistic for 4Bet hands only then that would require a new custom statistic with a specific test that the player in question made the 4Bet using the aggressors string. This post has information on how the actors and aggressors strings work while this thread and this thread discuss how you can compare/test the strings.

I looked at the stat definitions and 4bet ratio is using PFR% as the numerator and 4bet opportunities as the denominator while the latter is using RFI% as the numerator and 3bet defense opportunities as the denominator.

Using RFI is specific to open raises followed by 4Bets+ only however that's not going to include hands where the 2bettor raised limpers or 4Bet cold. If you want to count these types of hands as well they can be included in a new custom statistic.

More to the point, in what way 4bet opportunities produce different results than 3bet defense opportunities? Is it because you re facing all ins and you can't 4bet which artificially makes the 4bet range seem more narrow?

Yes that would be the case for for some hands.

[leviathan74]

Please note that this statistic in PokerTracker 4 is called '4Bet+ Ratio' and it is not specific to 4Bet hands only. The custom statistic called '4bet range (RFI * 4Bet)' in the Download Warehouse is also not specific to 4Bet hands only - it is 4Bet or higher. If you want a statistic for 4Bet hands only then that would require a new custom statistic with a specific test that the player in question made the 4Bet using the aggressors string. This post has information on how the actors and aggressors strings work while this thread and this thread discuss how you can compare/test the strings.

Thank you, I knew this, for the time being, it seems as if 5bets and higher are so few and far in between in my games that I don't think they bias this stat.

Although I do have to note that both stats are using the (cnt_p_4bet_after_raising) column and i don't think this variable includes 5bets and higher. Either way, it doesn't matter to me.

Using RFI is specific to open raises followed by 4Bets+ only however that's not going to include hands where the 2bettor raised limpers or 4Bet cold. If you want to count these types of hands as well they can be included in a new custom statistic.

Right, RFI is different than PFR, because PFR includes instances in which you raised limpers PF. My operating assumption -though I could be wrong - is that regardless of whether we are talking about a PFR or an RFI, people are going to 4bet+ the same range of hands more or less. Like, if you are going to 4bet KK+ - AK, it doesn't matter whether you were the PFR or RFI. I don't think there's any cold 4bets included in either stat though.

Yes that would be the case for for some hands.

So, I delved into a bit more and this what I see. When I filter for lower stack depths, such as 20BB or less, the 4bet+ ratio numbers are bigger than the RFI*4bet numbers. What explains this is that because the latter stat measures how often you 4bet when you were facing a 3bet and often times you are facing an all in -and thus can't 4bet- the 4bet numbers seem low. If I edit the stat to include 4bet opportunities instead of times you were defending against a 3bet, the stat gives higher numbers in lower stack depths.

What is perplexing to me now is that when I filter for high stack depths, let's say 95BB+, the RFI*4bet gives me larger numbers than 4bet ratio. For example, I am looking at a large sample size of the population and RFI*4bet tells me the pop 4bets 3% of the time whereas 4bet ratio tells me they are 4betting 2% of the time.

Which range should I think is accurate when I 3bet and face a 4bet? 3% or 2%?

[Flag_Hippo]

Although I do have to note that both stats are using the (cnt_p_4bet_after_raising) column and i don't think this variable includes 5bets and higher.

It can include some depending on the exact action.

Right, RFI is different than PFR, because PFR includes instances in which you raised limpers PF.

Yes although as long as a player raises anytime preflop it counts towards PFR and the first raise (or opportunity to raise) a player makes in a hand may not necessarily be a 2Bet.

What is perplexing to me now is that when I filter for high stack depths, let's say 95BB+, the RFI*4bet gives me larger numbers than 4bet ratio. For example, I am looking at a large sample size of the population and RFI*4bet tells me the pop 4bets 3% of the time whereas 4bet ratio tells me they are 4betting 2% of the time.

Which range should I think is accurate when I 3bet and face a 4bet? 3% or 2%?

That doesn't seem unreasonable. I'd expect the former to be higher and I would personally go with RFI*4Bet.

[leviathan74]

Yes although as long as a player raises anytime preflop it counts towards PFR and the first raise (or opportunity to raise) a player makes in a hand may not necessarily be a 2Bet.

What is perplexing to me now is that when I filter for high stack depths, let's say 95BB+, the RFI*4bet gives me larger numbers than 4bet ratio. For example, I am looking at a large sample size of the population and RFI*4bet tells me the pop 4bets 3% of the time whereas 4bet ratio tells me they are 4betting 2% of the time.

Which range should I think is accurate when I 3bet and face a 4bet? 3% or 2%?

That doesn't seem unreasonable. I'd expect the former to be higher and I would personally go with RFI*4Bet.[/quote]

Thank you. Delving even further into the stat, it seems to me that the discrepancy is caused by the fact that the PFR stat counts all raising bets, be it 2bet. 3bet, 4bet, 5bet etc. Obviously, if the PFR stat also counts 3bet+ bets, then these are hands in which a player can't 4 bet, so the resulting number seems smaller than it actually is. Am I reading this correctly?

And if I am, doesn't this make the 4bet+ stat flawed?

FWIW, I took the original formula for the 4bet * RFI stat:

(cnt_p_raise_first_in / cnt_p_open_opp)*100*(cnt_p_4bet_after_raising / cnt_p_3bet_def_opp_when_open_raised)

and changed it to :

(cnt_p_raise_first_in / cnt_p_open_opp)*100*(cnt_p_4bet_after_raising / cnt_p_4bet_opp_when_open_raised)

This seems to me to produce a more accurate number. Does this sound reasonable to you?