Aggression factor question

    • MIM1C
      MIM1C
      Bronze
      Joined: 24.05.2013 Posts: 18
      Hello. I am finally trying to set up my autorate rules in HM2 to distinguish different player types, and found myself in a tricky spot, the Aggression Factor stat. Can anyone tell me what AF values are standard for a NIT, TAG and a LAG? For 6max and Full ring separately (if this stat is different ofc).
      Thanks in advance!
  • 8 replies
    • AlvisR
      AlvisR
      Bronze
      Joined: 27.02.2009 Posts: 656
      what game?
    • DrDunne
      DrDunne
      Bronze
      Joined: 29.12.2010 Posts: 3,338
      assuming cash, i always just went with these values:
      0-1 = passive
      2 ~ somewhere in between
      3+ = aggressive

      however, i'm pretty sure AF is pretty unreliable. I stopped using it and replaced it with aggression frequency and W$WSF. these probably tell you a little more.
    • MIM1C
      MIM1C
      Bronze
      Joined: 24.05.2013 Posts: 18
      Originally posted by AlvisR
      what game?
      I am making rules for both Cash and MTT's/SnG's

      Originally posted by DrDunne
      assuming cash, i always just went with these values:
      0-1 = passive
      2 ~ somewhere in between
      3+ = aggressive

      however, i'm pretty sure AF is pretty unreliable. I stopped using it and replaced it with aggression frequency and W$WSF. these probably tell you a little more.
      Thanks for the numbers, I was thinking something similar.
      Although I'm sure W$WSF is a good stat to distinguish winning players from loosers, it will not help me define the player type. As far as I know, AF is most likely the best stat to tell how aggressive someone is post flop.
      So I just need someone with high poker experience to advise me that for example "passive players have an AF of 0-1, TAG's have at least 2, and typical LAG have it over 3 no matter what"

      I suppose cash is not that much different from MTT's, but if I'm wrong, please do tell me these values for both
    • VorpalF2F
      VorpalF2F
      Super Moderator
      Super Moderator
      Joined: 02.09.2010 Posts: 8,901
      Originally posted by MIM1C
      So I just need someone with high poker experience to advise me that for example "passive players have an AF of 0-1, TAG's have at least 2, and typical LAG have it over 3 no matter what"

      I suppose cash is not that much different from MTT's, but if I'm wrong, please do tell me these values for both
      Hi, MIM1C,
      I'm not sure "no matter what" applies here. There will be individual variations.

      You probably already know, but only 3 default autorate rules use Postflop aggression factor
      All 3 require >= 10 samples
      Loose Passive (mouse) AF < 2
      Loose Aggressive (dice) AF > 2
      TAG (eagle) AF > 2

      I'm incline to agree that an AF > 3 is more typical of a LAG esp 6-Max

      The default autorate rules all appear to be the same for all holdem games, so I can see why if you want to use Autorate, you need to edit them.
      A 6-Max TAG is much looser than a FR TAG, for example.

      Will you be creating some of your own categories?

      Best of luck with your project!
      --VS
    • MIM1C
      MIM1C
      Bronze
      Joined: 24.05.2013 Posts: 18
      Hello VorpalF2F, thank you for your reply!

      Originally posted by VorpalF2F
      Will you be creating some of your own categories?

      Best of luck with your project!
      --VS
      Yes, I am actually in the process of creating a multitude of player types: Scared NIT (low PFR, low WTSD), Rock, Tight Passive, ABC, ABC low 3bet, TAG, Lion (same range as TAG, but a lot more aggressive), LAG, LOOP Station, LOOP Fit/Fold, Maniac (Super LAG), Fish. This is why I need accurate ranges for AF, because that single stat will be a pretty big factor in distinguishing Lions from TAGs, Maniacs from LAG's and so on.
    • VorpalF2F
      VorpalF2F
      Super Moderator
      Super Moderator
      Joined: 02.09.2010 Posts: 8,901
      This sounds cool.
      Do you have a reference for typical stats for the player types?

      How do you handle the situation of type sets of stats overlap?

      Cheers,
      --VS
    • PerusJamppa
      PerusJamppa
      Bronze
      Joined: 18.01.2008 Posts: 2,714
      Originally posted by MIM1C
      As far as I know, AF is most likely the best stat to tell how aggressive someone is post flop.
      I have to disagree this a bit. Sure it's good enough for this purpose (and many others), but imo Agg% is slightly better stat when defining villains and their aggression.

      This is from HEM forum:

      b/b/b - AF: inf, AFq: 100, Agg Pct: 100
      b/x/b - AF: inf, AFq: 100, Agg Pct: 67
      b/x/x - AF: inf, AFq: 100, Agg Pct: 33
      b/x/c - AF: 1.0, AFq: 50, Agg Pct: 33
      b/c/c - AF: 0.5, AFq: 33, Agg Pct: 33
    • MIM1C
      MIM1C
      Bronze
      Joined: 24.05.2013 Posts: 18
      Originally posted by VorpalF2F
      This sounds cool.
      Do you have a reference for typical stats for the player types?
      Well, I did some reading on the topic and combined that with the default HM2 auto rate rules. Basically VPIP, PFR, 3bet% and Agg% is enough to distinguish player types pretty accurately.

      Originally posted by VorpalF2F
      How do you handle the situation of type sets of stats overlap?
      The HM2 Auto Rate Rule system has a priority list. If a player fits more profiles (for example someone with 41/32 might fit in the maniac category (assuming he has high aggression stats too) and the fish category), the player will be rated as whichever auto rate rule is higher. Obviously the auto rate rule for maniacs needs to be higher than the fish one.

      Originally posted by PerusJamppa
      Originally posted by MIM1C
      As far as I know, AF is most likely the best stat to tell how aggressive someone is post flop.
      I have to disagree this a bit. Sure it's good enough for this purpose (and many others), but imo Agg% is slightly better stat when defining villains and their aggression.

      This is from HEM forum:

      b/b/b - AF: inf, AFq: 100, Agg Pct: 100
      b/x/b - AF: inf, AFq: 100, Agg Pct: 67
      b/x/x - AF: inf, AFq: 100, Agg Pct: 33
      b/x/c - AF: 1.0, AFq: 50, Agg Pct: 33
      b/c/c - AF: 0.5, AFq: 33, Agg Pct: 33
      Hmm. Looks like you are right PerusJamppa! Agg% looks way more logical, and indeed is more reliable. Filtering Maniacs with 30+ VPIP, less than 1.8 V/P, and 3+ AF with at least 50 sample size returned no results, but changing the 3+ AF to an Agg% of 40+, gave me a good amount of players from my database with unusually maniacal stats to fit the profile.
      Thank you!