What is All-in EV in PT3?

    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      I heard that All-in EV is an rough indicator of how good you run. However I just don't understand how it is calculated. When there is an actual AI in a hand it makes sense. But when there is not, how the hell does PT3 calculate AI EV when it doesn't even know the opponent's hand?
  • 17 replies
    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      Or is it just based the starting hand vs. random hands? I guess not cause it doesn't make sense to ignore the board.
    • NightFrostaSS
      NightFrostaSS
      Bronze
      Joined: 25.10.2008 Posts: 5,255
      it does only calculate all in hands, that's why it's called ALL-IN EV
    • Jan217
      Jan217
      Silver
      Joined: 02.01.2009 Posts: 626
      whenever no one went all in before the river your all in ev is just exactly what happened, ie. if you won 5$ your 'all in ev' is +5$ and if you fold your big blind your 'all in ev' is -1bb.
    • Jan217
      Jan217
      Silver
      Joined: 02.01.2009 Posts: 626
      so really its actually a prety dire way of measuring whether or not you're running good. :p
    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      So if I am a 2-1 underdog going AI on flop/turn and make a suck-out, what's the AI-EV going to be for this hand? Is it going to be 33% x the pot size?
    • tokyoaces
      tokyoaces
      Bronze
      Joined: 01.04.2009 Posts: 1,883
      Originally posted by yeahyoung0312
      I heard that All-in EV is an rough indicator of how good you run.
      Who ever told you this was misinformed. It only shows your EV from the point that you went all-in at.

      To me its more of an indicator of how good or bad your all-in decisions are skewed by the size of the individual pots that those decisions apply to. So I guess its an indicator of how poorly you play as the pots get big? :f_confused:
    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      Originally posted by tokyoaces
      To me its more of an indicator of how good or bad your all-in decisions are skewed by the size of the individual pots that those decisions apply to. So I guess its an indicator of how poorly you play as the pots get big? :f_confused:
      I see. So when I go in AI as an underdog my AI-ev for that hand will be negative but if I make a suck-out then my showdown winning for that hand will be highly positive.

      How about non-AI hands? Is it as Jan217 said:

      Originally posted by Jan217
      whenever no one went all in before the river your all in ev is just exactly what happened, ie. if you won 5$ your 'all in ev' is +5$ and if you fold your big blind your 'all in ev' is -1bb.
    • nitaidean
      nitaidean
      Bronze
      Joined: 15.05.2010 Posts: 60
      There's a shortcoming to the All-In EV in my opinion... say I go "almost" all-in on the turn, with just 5c left to bet, and I go in with 85% to win. And say he hits his backdoor flush on the river and sucks me out, and bets the additional 5c, which I naturally call. Now in this case, my All-In EV diff is zero. Meaning, because I wasn't technically "all-in", it doesn't count that I was sucked out.

      I think it should be different: all hands that make it to showdown, the program should then go back in the hand, and on every street it calculates my equity and the opponent's equity, and in this way adjusts our winnings or loses for luck.

      For instance, say I have AA and my opponent KK.

      Preflop: I bet $1 and he call, so I'm now +$0.84c and he's +$0.84c, because thats our equity preflop.

      Flop (7 9 2): I bet $1.50 and he calls, so now I get an additional +$1.30 and he gets -$1.30.

      Turn (K): I bet$4.00, he reraises all in to $15.00 and I call. Now I'm +$0.6 or so (my only equity is hitting an A on the river) and he is +$14.40.

      So here I get credit for the streets I was ahead, and lose the streets I was behind.
    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      Originally posted by nitaidean
      Turn (K): I bet$4.00, he reraises all in to $15.00 and I call. Now I'm +$0.6 or so (my only equity is hitting an A on the river) and he is +$14.40.
      My understanding is equity is not EV. And IMO in this case your ev is roughly $30 x 4% + (-$15) x 96% = - $13.2
    • tokyoaces
      tokyoaces
      Bronze
      Joined: 01.04.2009 Posts: 1,883
      Originally posted by nitaidean
      There's a shortcoming to the All-In EV in my opinion...
      You suggestion would not make it any better. It would just be a different but equally flawed system.

      The problem with all-in EV is that it completely ignores all of the decisions leading up to the point that both players decided to go all-in. The decision point is neither pre-flop nor at the point of the effective all-in.

      However coming up with a real solution, that is checking the EV against the villain's likely range, is too hard because we don't have infinite computing power. So instead someone decided just to take the easy way out.

      I'm convinced that the only reason they did this is because it sells more copies of tracking software.
    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      Originally posted by tokyoaces

      However coming up with a real solution, that is checking the EV against the villain's likely range, is too hard because we don't have infinite computing power. So instead someone decided just to take the easy way out.

      I'm convinced that the only reason they did this is because it sells more copies of tracking software.
      Come on. You must have seen plenty of flawed stats going on the research/applied fields.

      Anyway players need something to indicate their "card luck".
    • tokyoaces
      tokyoaces
      Bronze
      Joined: 01.04.2009 Posts: 1,883
      Originally posted by yeahyoung0312
      Come on. You must have seen plenty of flawed stats going on the research/applied fields.
      I believe that is called "Marketing". :s_biggrin:

      Originally posted by yeahyoung0312Anyway players need something to indicate their "card luck".
      Yes, it appears so. :f_cool: I guess that makes the creator a marketing genius! :f_p:
    • dadaas
      dadaas
      Bronze
      Joined: 16.03.2010 Posts: 88
      ALL IN EV is good indicator, it will show you how unlucky or Lucky are you. Or how good or bad player you are. IMO if you are running above EV then you must consider your strategy because you are most probably fish and soon you will be hit by variance and then you could lose all money you have, depends how much you want to gamble.

      For example, i m playing SSS and its imposible to run above EV ALL IN, because you rarely need to sockout and most of the time if you lose you will be bad beat.

      So yes this indicator is weird and it should be taken into consideration... As example above... Where i play only good hands and i dont need to sockout... and if i lose i m heavy bad beat which will move EV line ...
    • yeahyoung0312
      yeahyoung0312
      Bronze
      Joined: 16.12.2009 Posts: 340
      Originally posted by dadaas
      ALL IN EV is good indicator, it will show you how unlucky or Lucky are you. Or how good or bad player you are. IMO if you are running above EV then you must consider your strategy because you are most probably fish and soon you will be hit by variance and then you could lose all money you have, depends how much you want to gamble.
      I'm confused. Why do I need to adjust when I am running hot?
    • Gu911
      Gu911
      Bronze
      Joined: 03.05.2010 Posts: 987
      IMO the biggest problem with All-in EV is that i can differ greatly over a small sample. Say you get it in with AA against preflop with effective stacks of 100bb. Your Equity here is 81-82% depending on suits. You win the hand and pick up the pott of 200bb. The EV-line shows a profit of 82bb, but your line for SD-winnings shows a 100bb profit.
    • pzhon
      pzhon
      Bronze
      Joined: 17.06.2010 Posts: 1,151
      Originally posted by Gu911
      IMO the biggest problem with All-in EV is that i can differ greatly over a small sample. Say you get it in with AA against preflop with effective stacks of 100bb. Your Equity here is 81-82% depending on suits. You win the hand and pick up the pott of 200bb. The EV-line shows a profit of 82bb, but your line for SD-winnings shows a 100bb profit.
      Actually, if you get all-in with 82% equity in a 200 bb pot, your luck-adjusted equity for the hand will be 64 bb = 82% of 200 bb - 100 bb. The 64 bb ignores whether you won or lost. This should be a better unbiased indicator of your skill advantage than the results of +100bb or -100bb.

      Luck adjustment does not have to correct for all possible sources of luck. If you could filter out all of the luck you have on Fridays, leaving the luck from the other 6 days of the week, this would be an improvement. When you play cash games, the all-in luck adjustment helps you to get a more accurate estimate of your win rate even though it does not factor out coolers, whether you hit the flop, or other types of luck.

      All-in luck adjustment is much more effective if you play in games where you often get all-in preflop or on the flop. This can occur if you play with a short stack yourself or against short stacks, or if you play tournaments, particularly super turbo SNGs where players start with 10 bb stacks. All-in luck accounts for about 73% of the variance in super turbos, so the adjusted results after 270 tournaments are about as reliable as the raw results after 1000 tournaments. In a big stack cash game, however, only 10-20% of the variance might be eliminated by using this luck adjustment.
    • MiiWiin
      MiiWiin
      Moderator
      Moderator
      Joined: 01.03.2007 Posts: 64,801
      You should be very careful with this all-in EV.

      But there is no other chance to make such an analysis, it is not possible to examine the "EVs" on the other streets.

      If it theoretically works, you just have the showdown-EV. But what is about Bluffs? Semi-Bluffs? And the resultant foldequity?

      What about a raise with 56s preflop? Always -EV when youre just looking for the equity.

      A 3-barrel-bluff...what about flop and turn? Maybe 0% Equity, but you win the pot at the river.

      There is no other way to make such a EV-line.