What the hell? Way below EV on a huge sample

    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      This is what happened :

      I changed my nickname and started playing BSS at NL10 after playing 90k hands or so MSS. I changed the strat because I wasn't happy with the winrate I was having, plus that I learned enough to play BSS. After 2 days I had incredible badbeats and you can see my graph :



      But that is super small sample so I thought "Hey, let's check my old graphic", by the way I recently downloaded HEM free trial, I was using a freeware hud before, so I never had the chance to check my graphics. This is it :



      Keep in mind that I was playing with 40bb stacks the whole time, that's -10 BI over a huge, I dare to say, sample. Can anyone enlighten me if this is normal?

      My EV adjusted winnings are 383.69$ but my true winnings are 346.07$.
  • 21 replies
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,297
      If your EV line had been shooting back and forth across your green line +10/-10 etc, you'd think nothing of being 10bi above EV.

      This is perfectly normal, and 100k is perhaps not as large a sample as you think it is (variance is huge, bigger than most players actually imagine). It's not unheard of to be 40bi below ev over 100k hands for example.
    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      So you mean that a cumulated -80$ difference over 340$ winnings is nothing?
      Why do I never run above the EV line? :(
    • Elroch
      Elroch
      Bronze
      Joined: 08.02.2011 Posts: 150
      @BadeaCelRau, rather than wondering whether this sort of difference is unusual, why not do the easy calculation to see how significant it is?

      Suppose you have a standard deviation of 80bb/100. Then on 100,000 hands, the SD is 80*sqrt(1000)bb ~= 2500 bb. The difference on your graph is 2000 bb. About 25% of people have a worse difference than yours making it entirely unsurprising.

      Of course more precise numbers will give a slightly different result, but a similar conclusion.
    • OZSA
      OZSA
      Bronze
      Joined: 18.05.2009 Posts: 804
      sometimes I have 10BI EV difference in one day..wish I had 80$ ev difference in 120k hands...wow
    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      Who established the standard deviation of 80bb/100 ?(

      I'll check all the graphs I see posted here and see if anyone has a worse graphic.

      Originally posted by OZSA
      sometimes I have 10BI EV difference in one day..wish I had 80$ ev difference in 120k hands...wow
      Dude that's 1/4 of my winnings ... I don't know what limit are you playing
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,297
      Originally posted by BadeaCelRau
      So you mean that a cumulated -80$ difference over 340$ winnings is nothing?
      Why do I never run above the EV line? :(
      Yup....unfortunately.

      It's better to come to terms with this fact sooner rather than later. Some people have even quit their day jobs before realising what variance can do to them. :f_cool:
    • ihufa
      ihufa
      Gold
      Joined: 18.03.2008 Posts: 3,323
      LOL THAT'S NOTHING I'VE BEEN 10 BUYINS BELOW EV IN A SINGLE SESSION
    • ihufa
      ihufa
      Gold
      Joined: 18.03.2008 Posts: 3,323
      MORE THAN ONCE
    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      I cheked the first 6 pages from the bad graphs topic and I only found PLO graphs, or very few hands until this guy called cryoburn posted his Party graph

      http://j.imagehost.org/view/0406/mega_underEV

      PartyPoker = thieves?

      @ihufa
      Wanna see my 888 graphic? The 2 lines got nothing in common but the results are very close in the end. I'm talking about a big sample, not a single session.
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,297
      Just thought I'd share this with you guys. Sorry, I cut off x-axis, but it's basically 1-100k sample. (multiply x-axis by 100 for number of hands)

      It's from a variance simulator I downloaded, where you input a std. deviation and a winrate. You can then repeatedly press F9 to generate some different graphs you might expect given those two factors.

      Most are positive given a positive winrate. Some however, are shocking.....



      Uploaded with ImageShack.us
    • Snodreliss
      Snodreliss
      Silver
      Joined: 23.12.2007 Posts: 1,280
      BadeaCelRau, sorry to say it, but that's not really a big sample :( Look at it in a good way, you are still winning even though you are abit under EV ;)

      GL
    • Edukator99
      Edukator99
      Bronze
      Joined: 18.07.2010 Posts: 240
      Am I the only one who just doesn't think this stat matters? As far as I understand it's only measured when you go all in BEFORE the flop. So it's a measurement of how you fare when premium hands match up. I mean...what does that tell you about the rest of your play? Say you only go in with AA or KK...well you'll probably do well... I guess as you're only getting expected EV measured on your top 2 hands.

      What if you extend that to AA, KK, QQ, JJ, 10 10, , 99AK suited, AKo, AQs, AQo. I mean...you're going to have a pretty tough time matching the expected EV line of the first guy.

      I just don't get how it really reflects on your ability. All in preflop is what? A few % of your hands?

      Maybe someone could clarify this..
    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      @w34z3l
      Cand you please remind me the name of that site?

      Anyway I lost faith in PartyPoker. Nobody plays there anyway and after seeing the graphs :/ ... I don't see myself running above EV too soon.
    • w34z3l
      w34z3l
      Coach
      Coach
      Joined: 03.08.2009 Posts: 13,297
      @BadeaCelRau

      Head over to anskypoker.com and look for it in the "tools" section.

      @Edukator99

      Allin-ev is calculated from anytime you are allin, regardless of it being preflop or postflop. Anytime you do not go allin, your allin-ev equals the amount won or lost in the hand.

      You are correct in that it is possibly not that useful; it just measures one particular type of variance, and is not representative of your true expected value.
    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      So how many more hands should I keep playing to be above EV? :)

      Hand converted with online PokerStrategy.com hand converter:

      Play hand

      $0.05/$0.1 No-Limit Hold'em (8 handed)

      Known players:
      UTG2:
      $10.28
      MP1:
      $0.10
      MP2:
      $4.40
      MP3:
      $10.50
      CO (Hero):
      $10.00
      BU:
      $9.69
      SB:
      $9.50
      BB:
      $3.00


      Preflop: Hero is CO with K, K.
      Hero posts deadblind, 4 folds, Hero raises to $0.40, BU folds, SB calls $0.35, BB folds.

      Flop: ($0.9) 4, 9, T (2 players)
      SB bets $0.10, Hero raises to $0.90, SB calls $0.80.

      Turn: ($2.7) 7 (2 players)
      SB checks, Hero bets $2.57, SB calls $2.57.

      River: ($7.84) 6 (2 players)
      SB bets $0.10, Hero calls $0.10.

      Final Pot: $8.04.
      Results follow:

      Hero shows a pair of kings(K K).
      SB shows a flush, ten high(4 2).

      SB wins with a flush, ten high(4 2).



      Hand converted with online PokerStrategy.com hand converter:

      Play hand

      $0.05/$0.1 No-Limit Hold'em (8 handed)

      Known players:
      CO (Hero):
      $10.25
      BU:
      $11.98
      SB:
      $2.29
      BB:
      $3.25
      UTG2:
      $10.15
      MP1:
      $10.10
      MP2:
      $10.67
      MP3:
      $4.85


      Preflop: Hero is CO with Q, Q.
      4 folds, Hero raises to $0.40, BU calls $0.40, SB raises to $2.29, BB folds, Hero calls $1.89, BU folds.

      Flop: ($5.08) 7, 2, T (2 players)


      Turn: ($5.08) 9 (2 players)


      River: ($5.08) 4 (2 players)


      Final Pot: $5.08.
      Results follow:

      SB shows three of a kind, tens(T T).
      Hero shows a pair of queens(Q Q).

      SB wins with three of a kind, tens(T T).
    • randomdonk
      randomdonk
      Bronze
      Joined: 02.08.2009 Posts: 1,984
      [ ] way below EV
      [ ] huge sample
      [x] EV line is useless
    • patszerdonk
      patszerdonk
      Bronze
      Joined: 19.05.2011 Posts: 834
      There is only 1 bad beat at around hand #1000. And small sample.
      It's normal

      and (if I understand it correctly) it's very hard to run above AI EV, if you only all in with KK and AA or other premium cards.

      If you want to run above AI EV, try this:

      everytime you get 72o, go all in preflop! :f_biggrin:
    • BadeaCelRau
      BadeaCelRau
      Bronze
      Joined: 15.10.2010 Posts: 2,122
      I checked my NL4 graph and it appears it is possible to run above EV :-O



      @patszerdonk
      That's not true, the line is calculated every time you are all-in or at showdown, ideally it should match your winnings perfectly. But most graphs from winnings players I've seen run above EV (but nobody mentions being on an upswing).

      Maybe I'm super lucky at picking up lucky cards when calling with the worse hand :D and that compensates for a below EV graph. Anyway, you guys convinced me, sometimes the fishes donate more than your expected gain. I'll keep playing a few hands and post here when I run above EV (currently 360bb under :f_frown: )
    • warmedia
      warmedia
      Bronze
      Joined: 25.11.2008 Posts: 149


      in 1 session

      trust me its easy to get under ev
    • 1
    • 2