Usefullness of AF (or lack thereof)

I think I've said here before, I don't even have AF show up on my PAHUD, I exclusively use the aggression percentage. With the AF someone who raises or bets 0.5% of the time and folds the other 99.5% of the time has infinite aggression. What do you use this number for? From what I've heard mostly it's for how likely they are to be bluffing, so if they bet with a high AF it's likely they are bluffing whereas with a small AF they most likely have the goods. I just don't understand or like counting folds as neutral and calls as anti-aggressive. Am I missing something? Why is AF one of the standard 3 stats to define any given opponent with a flaw like that?

I think you're right about it helping determine someone's bluffing frequency. Or semi-bluffing as well, if someone with an AF<1 raises a wet flop I tend not to put them on a draw. But I agree with you about how misleading the infinity stat is. Without any other solid reads about whether an opponent is passive or aggressive, I usually disregard an infinity AF stat until I've played enough hands with them for it to reach a more accurate ratio.

This is a very good question and I'm also interested in others' opinions on this.

I prefer bet% stats than aggression stats.

Like knowing that someone bets 86% of their rivers means that you can much more easily extract value by c/r your monster on the river. The AF makes no sence as some players have infinite and others who are also considered agressive have 4.3 or something.

As for the standard stats it gives you a general idea across all areas how agressive they are, the higher the more aggressive but I never really understood it. i mean how high does it go before its infinity?

Well my example was just that, an example. If someone bets 10%, calls 0.5%, and folds 90.5% they have a 20 AF, it's not just infinity. And it comes up with infinity when you get a positive number divided by 0.

But my comment wasn't about results anyway, I'm just talking about the methodology behind it. Why would a fold be considered neutral but a call unaggressive? I don't care what results it comes up with (well I do, but not for now), I just was using examples to show the flawed methodology, treating a fold as neither aggressive nor unaggressive when most would think it unaggressive, and a call the most unaggressive move of all. The aggression % treats both folds and calls as unaggressive and I like that stat much more than the traditional AF. I'm just wondering why AF is considered one of the major stats used by pretty much everyone to describe a player. Why isn't aggression % used instead?

whats the equation for AF?

I think that it takes into account aggression % or at least bet %.

Will try to find it on 2+2 :/

It's something like the ratio of (bets+calls) : (checks+folds) I think. I could be completely wrong.

Anyways I've often thought about this too, zach. I've heard from numerous people (most of them a lot better than us) that AF is the most 'useless' stat. As of right now I don't bank my decisions on it but am trying to figure out exactly when to use it and how (if at all), and also for aggression frequency.

AF = (# of times bets + # of times raises) / (# of times calls). All post-flop, I believe. Folds do NOT factor in. If you have a small number of hands on an opponent who has bet and/or raised but never called post-flop, their AF would be infinity. This is misleading at first, because they could be an aggro maniac, always betting/raising, or they could be timid, always folding but sometimes betting strong hands.

AF is important to help you decide when you should be c/cing every street with something like TPGK, and generally when one should be slowplaying.

ie: if you open AQ in MP and guy with high AF calls from button, you could easily get the most value out of your hand by check-calling down on A8366 board

Well true, but what we're debating is how much we can rely on the stat since it can be easily skewed.

fyi there has been a suggestion in the PT3 forums about aggression factor being revamped to include bet sizing. Very interesting point and it would be SO much more useful if that were true...