Aggression formula

[Luke Montgomery]

If the formula is BET%+RAISE%/CALL% what is the highest number produced by this.

My aggression factor is 1.4 but I guess i'm trying to figure out the scale of Min, Max, and avg.. any ideas?

[uuDevil]

[ QUOTE ]

If the formula is BET%+RAISE%/CALL% what is the highest number produced by this.

My aggression factor is 1.4 but I guess i'm trying to figure out the scale of Min, Max, and avg.. any ideas?

[/ QUOTE ]
No reason it couldn't be zero if there are players who never lead or raise but will call bets or raises. No reason it couldn't be infinite if there are players who will bet or raise but never call. What the distribution may be you'd have to determine from collected data. That might be a fun exercise to do with your PokerTracker data for other players.

[Siegmund]

Yes, AF can range from 0 to infinity.

I am a bit puzzled why the PT folks chose to define it this way which makes it a bit of a slippery statistic to work with.

If you are interested in confidence intervals or convergence rates, you may find it easier to work with AF/(1+AF), which is a garden variety binomial p confined to [0,1] like VPIP and PFR are (but with "number of actions on this street" rather than "number of hands" for n.)

If t'were my program, actually, I'd count # of hands on which I took an agressive action on a given street and divide by total # hands where I took any action on the street. (That is, count a limp-reraise as "one agressive action" and a call-1-and-then-call-another or a call-1-and-then-fold-to-a-bet as "one passive action", rather than counting individual calls and raises.)