All-In-EV

    • Smileyphil
      Smileyphil
      Bronze
      Joined: 28.10.2008 Posts: 488
      I always seem to be below the suggested value for EV on Holdem Manager.

      I am beginning to wonder how accurate it is.
      I understand that it is not a representation of "good play". It is possible to have done exactly the right thing and have negative All-In-EV for example if you shove KK and are called by AA.

      I am instead talking about over larger samples (at least 20k but preferably a lot more). Over a large sample the situational events (coolers and the like) will balance out and therefore we should be able to view a positive EV as a sign that we have a strong/winning game?

      However my data shows that the All-In-EV is far greater than my actual winnings all the way along. I'm either very unlucky or it simply calculates something wrong?

      I play SSS so I am generally all in preflop or on the flop and therefore I look to the All-In-EV of a session (or in total) as a feel of how well I am playing.

      Does anyone have any comments or information about this?
  • 7 replies
    • jonnyjm
      jonnyjm
      Bronze
      Joined: 24.03.2008 Posts: 447
      Originally posted by Smileyphil
      I always seem to be below the suggested value for EV on Holdem Manager.

      I am beginning to wonder how accurate it is.
      I understand that it is not a representation of "good play". It is possible to have done exactly the right thing and have negative All-In-EV for example if you shove KK and are called by AA.

      I am instead talking about over larger samples (at least 20k but preferably a lot more). Over a large sample the situational events (coolers and the like) will balance out and therefore we should be able to view a positive EV as a sign that we have a strong/winning game?

      However my data shows that the All-In-EV is far greater than my actual winnings all the way along. I'm either very unlucky or it simply calculates something wrong?

      I play SSS so I am generally all in preflop or on the flop and therefore I look to the All-In-EV of a session (or in total) as a feel of how well I am playing.

      Does anyone have any comments or information about this?
      We'll 20k is still quite a small sarmple size tbh...things will be/should be more accurate after 50k hands.

      Also though say you shove AA on the flop with 60BB infront of you, You get called by somone who has KK or QQ with only 40BB infront of them...This doesnt count at all-in EV as you still have money in front of you (atleast this is the case with pokertracker (stupid i know)). So if they suck out on you it will be shown as negative ev to 100% of the $$$ you lost in the pot...the same if you win but positive.
    • LuborC
      LuborC
      Bronze
      Joined: 20.04.2008 Posts: 1,243
      20K is waaaaay to few hands. It is true that everything should even out in the long run but that mean rather millions than thousands of hands. It is of course possible to run exactly as well as expected after 10 hands you played but it's also possible to be way off after 100 mil hands.
      Generally speaking only the probability that you come close to the EV increases with the number of hands but you can never be certain that you will ever reach it. Unless you are able to play infinite amount of hands that is..
      I for once also use HEM and I have never really dropped under the EV in the 25K hands I've played so far. But it does not really matter as there is obviously nothing you can do to influence this. You just have to take the variance as it comes.
    • Smileyphil
      Smileyphil
      Bronze
      Joined: 28.10.2008 Posts: 488
      Whoa surely that can't be true about Pokertracker. That'd throw the whole idea of All-In-EV out completely?

      I see that it won't be exact but can I feel happy if I end a (large) session with + All In EV but negative winnings?

      I find in practice that most of my sessions end with +All-in-EV but not always with +winnings.

      Will our winnings tend towards our All-in-EV the more hands we play?
    • Meiffert
      Meiffert
      Bronze
      Joined: 13.10.2008 Posts: 151
      Originally posted by Smileyphil
      Will our winnings tend towards our All-in-EV the more hands we play?
      I'm not sure about it.
      Your style can actually have some influence on whether you run above or under EV in my opinion.
      I actually have quite a few hands which look like this:

      ***** Hand History for Game 31862278170 ***** (Poker Stars)
      $100.00 USD NL Texas Hold'em - Friday, August 21, 12:42:38 ET 2009
      Table Aspasia IV (Real Money)
      Seat 5 is the button
      Seat 1: 1ee7 ( $127.65 USD )
      Seat 3: 24CMSLAK ( $127.70 USD )
      Seat 4: Meiffert ( $20.00 USD )
      Seat 5: GKallai ( $58.20 USD )
      Seat 6: johnphil82 ( $103.25 USD )
      Seat 7: hilario1 ( $23.40 USD )
      Seat 8: RookieBookie ( $21.50 USD )
      Seat 9: stefir ( $17.00 USD )
      johnphil82 posts small blind [$0.50 USD].
      hilario1 posts big blind [$1.00 USD].
      ** Dealing down cards **
      Dealt to Meiffert [ Kh As ]
      RookieBookie folds
      stefir folds
      1ee7 folds
      24CMSLAK folds
      Meiffert raises [$3.00 USD]
      GKallai folds
      johnphil82 folds
      hilario1 calls [$2.00 USD]
      ** Dealing Flop ** [ 7d, Tc, Kc ]
      hilario1 checks
      Meiffert bets [$4.50 USD]
      hilario1 calls [$4.50 USD]
      ** Dealing Turn ** [ Ts ]
      hilario1 checks
      Meiffert bets [$12.50 USD]
      hilario1 calls [$12.50 USD]
      ** Dealing River ** [ Jc ]
      hilario1 shows [Td, 4c ]
      Meiffert shows [Kh, As ]
      hilario1 wins $38.55 USD from main pot



      ***** Hand History for Game 31813954569 ***** (Poker Stars)
      $100.00 USD NL Texas Hold'em - Wednesday, August 19, 08:00:04 ET 2009
      Table Palma (Real Money)
      Seat 7 is the button
      Seat 1: 24CMSLAK ( $100.00 USD )
      Seat 2: SleepyWill ( $61.50 USD )
      Seat 3: Anand19 ( $20.00 USD )
      Seat 4: blackietje ( $21.20 USD )
      Seat 5: WojtasPolak ( $100.00 USD )
      Seat 6: Saschn ( $15.70 USD )
      Seat 7: serua ( $23.00 USD )
      Seat 8: Ventrue2 ( $25.00 USD )
      Seat 9: Meiffert ( $19.50 USD )
      Ventrue2 posts small blind [$0.50 USD].
      Meiffert posts big blind [$1.00 USD].
      ** Dealing down cards **
      Dealt to Meiffert [ Qd Ah ]
      24CMSLAK folds
      SleepyWill calls [$1.00 USD]
      Anand19 folds
      blackietje folds
      WojtasPolak folds
      Saschn folds
      serua folds
      Ventrue2 folds
      Meiffert raises [$4.00 USD]
      SleepyWill calls [$4.00 USD]
      ** Dealing Flop ** [ 9h, Qh, Ts ]
      Meiffert bets [$6.00 USD]
      SleepyWill calls [$6.00 USD]
      ** Dealing Turn ** [ 8s ]
      Meiffert bets [$8.50 USD]
      SleepyWill calls [$8.50 USD]
      ** Dealing River ** [ 5c ]
      Meiffert shows [Qd, Ah ]
      SleepyWill shows [Ad, Jd ]
      SleepyWill wins $37.60 USD from main pot


      In another hand I had a set and didn't quite move the slider while shoving the turn and only bet 10 bb out of my 10.5 bb stack. :-)
      The villain called, hit his flush on the river and my "All-in EV" in this hand was - 20 bb, altough in reality it should have been about ~ +16 bb.

      I think that if you play SSS, you usually have the stronger hand preflop (as you raise with a stron range) and your opponents are more often drawing with suited connectors and stuff like that. Therefore if they outdraw you on the turn/river when you only have half pot left in your stack, you stick it in drawing dead or very slim altough you were decent favourite both preflop and on the flop when most of the money went in and you were already decided that the rest is going in on the turn no matter what with your TPTK/overpair or something like that.
      I'm therefore kinda surprised that you actually run under your EV (maybe it's just a variance and you have been unlucky). I would like to see how you play. :-)
    • LuborC
      LuborC
      Bronze
      Joined: 20.04.2008 Posts: 1,243
      It only makes sense to consider ALL-IN EV if you get ALL-IN with some cards still to come. The EV stat tries to take the short term luck out of the picture by showing the average result after the cards are dealt.
      In theory it could also calculate the average result of every hand where you got to see the opponents' cards but it doesn't work that way. There would still be times where you didn't get to see the cards so it would not be accurate anyway.

      In other words ALL-IN EV shows you that you should have won those 20% of the times when you ran into AA with your KK and adds those 20% of the pot to your expected winning. It sets right the fact that you got a bit unlucky and lost 100% of the hand even though you still had 20% chance to win. It takes into account that you should still win 20% of the times when this happens and therefore shows you how this will end on average/ in the long run.
      What it does not take into account is the fact that you already got unlucky when this whole situation occured.

      So the hands that Meiffert posted do not change your ALL-IN EV because there was no more luck involved after the money went in. How lucky or unlucky you got before the money went in it does not take into account.


      So to answer your original question: ALL-IN EV shows somewhat more precise picture of how your session should have turned out but not a perfect one.
    • Meiffert
      Meiffert
      Bronze
      Joined: 13.10.2008 Posts: 151
      Originally posted by LuborC
      So the hands that Meiffert posted do not change your ALL-IN EV because there was no more luck involved after the money went in. How lucky or unlucky you got before the money went in it does not take into account.
      Well, they don't change the all-in EV compared to winnings, but they change it compared to not playing the hands at all (unless I understand it wrong).
    • Kruppe
      Kruppe
      Black
      Joined: 20.02.2008 Posts: 2,144
      to find out how accurate the all-in EV line in HEM is, just look for a hand where that line differs from the winnings line and look at it in the replayer, where you can see the percentages that lead to the line. you'll see that it's pretty simple, and you can check with stove or whatever

      it's called 'all-in EV', but it isn't al-in EV, it's just an all-in hand vs hand equity-adjusted winnings line.