Wtsd & W$Sd

    • JLeitmotiv
      JLeitmotiv
      Bronze
      Joined: 02.01.2009 Posts: 756
      I have the following quite-theoretical question. Suppose someone has these stats in 6max:


      Saw Flop: 30%
      WTSD: 49
      W$SD: 60
      Winrate: 4BB/100

      Definitely he's in godmode. However, he definitely won't keep these stats for too long. Is there anyway we could realize his true winrate when he W$SD goes close to the typical value for that WTSD (let's say, about 40)?

      What if he usually plays WTSD 38 W$SD 55? Can we somehow estimate his true winrate?
  • 8 replies
    • madorjan
      madorjan
      Bronze
      Joined: 13.11.2009 Posts: 5,561
      Hi there,

      The answer is simply no.:)

      The problem with these stats is they don't represent the potsize in which he won or lost. Let's assume a very extreme scenario: a player plays in an environment, where everybody else except for him can only check if checked to, and call if bet into. Therefore our player will only bet, if he's a statistical favourite against random hands - therefore making the potsize a whole lot bigger, than if any other player has the best hand.

      His WTSD will be a 100%, his W$SD will be really low (1/6 = 16.7%), however he's going to be a huge winner in these games, with seemingly awfully fishy stats. And we also have to add, that our player plays perfectly, it's just that the environment forces him to have these stats, since they are the one who never fold, and pay him off with every kind of trash.

      Now of course this example is quite extreme, but certainly is kind of close to what microlimit regulars can expect. I remember when I was playing the micros, and one of my coaches asked if I go to showdown too much - I said hopefully not, they're probably just coming from the limped pots, and the nature of the loose-passive games, and as far as I can tell, this was the case.

      Well, maybe not, I just looked at these stats for the current year, I'm winning with a decent winrate, have a WTSD of 45 and a W$SD of 49!! I think it's just the matter of putting enough money into the pot, when you have the best of it (thin vbetting and stuff).

      If you want to estimate his winrate through statistical analysis, take a look at here: http://www.evplusplus.com/poker_tools/poker_winrate_confidence/ - this handy tool will tell you, given your sample size, actual winrate and standard deviation, on what range his possible winrate is with a given confidence.

      Hope I could help,
      Madi
    • YohanN7
      YohanN7
      Bronze
      Joined: 15.06.2009 Posts: 4,086
      A W$SD of 60% looks awsome at first glance. In reality it's truly awful. Such a player folds much much much too much and probably bluffs too little.

      My own stats are, I dare say, close to "perfect" whatever that is. I'm smack dab in the middle of any reasonable recommended stats there is.

      But, alas, the past half year or so, I've been completely done over by players having anything but sound stats.

      Ironically, there is one stat for me that is on the borderline of being sound. It's the W$SD. It's a tad low. I choose to interpret this as saying that I'm inherently unlucky because there is absolutely nothing I can do statistically to change that nmber without ruining the others:D

      [The real problem, of course, is that I don't get my money in good as much as I used to.]

      /Johan = :f_confused:
    • JLeitmotiv
      JLeitmotiv
      Bronze
      Joined: 02.01.2009 Posts: 756
      Originally posted by YohanN7
      A W$SD of 60% looks awsome at first glance. In reality it's truly awful. Such a player folds much much much too much and probably bluffs too little.

      My own stats are, I dare say, close to "perfect" whatever that is. I'm smack dab in the middle of any reasonable recommended stats there is.

      But, alas, the past half year or so, I've been completely done over by players having anything but sound stats.

      Ironically, there is one stat for me that is on the borderline of being sound. It's the W$SD. It's a tad low. I choose to interpret this as saying that I'm inherently unlucky because there is absolutely nothing I can do statistically to change that nmber without ruining the others:D

      [The real problem, of course, is that I don't get my money in good as much as I used to.]

      /Johan = :f_confused:
      Well, in my opinion a high W$SD doesn't mean you are folding too much. That's having a low WTSD. For instance, in my HUD i have the WTSD and W$SD of the opponents, not because W$SD is inherently significant, but because it gives a light on the WTSD. If a player has a WTSD of 50 and a W$SD of 55, it doesn't necessarily mean he doesn't fold enough hands or that he doesn't bluff enough. It can easily mean that he's been having a pretty nice upswing in the hands I have seen so far.

      I don't think W$SD has any meaning by itself at all, but it helps you understand better the WTSD of the opponents and make better decisions.
    • YohanN7
      YohanN7
      Bronze
      Joined: 15.06.2009 Posts: 4,086
      The W$SD is one of the more important stats when analyzing your own game. Do you have Super System 2? Mike Caro explains this very well there.

      If you have a high W$SD, then adding a few more hands to call with, will naturally lower the W$SD. The reason is of course that you will lose the majority of these extra hands. But that is ok, since the few extra pots you win will more than compensate for the losses of one extra BB in the pots you lose. (But this will not necessarily be the case if you have a W$SD of 45%.) The simple explanation is that 1BB<<1POT. If you lose 80% of the pots when you call with [I]the bottom[/I] of you range, then you call too little because normally 1POT:1BB > 4:1.

      What I'm saying is that if you have a persistent W$SD of 60%, then you are too tight. Preflop, flop, turn or river, doesn't matter when you are too tight. You fold too much period. Whether you bluff too little is less obvious, but the same principles apply.

      Of course, there is a correlation between WTSD and W$SD, that's obvious. But when you look at your own numbers, then consider W$SD first, because this number has the direct calculable impact I outlined above. The WTSD is much less a direct indicative. It depends (should depend) a lot on how tight you are preflop for instance. When you look at [B]your opponents[/B] stats though, WTSD is much more important than W$SD. (Should you bluff him? Does he see the river with bottom pair/A high?)

      If you don't believe me, consider the pathological examples of a W$SD of 100% and (say) 25%. The optimal value of W$SD lies somewhere in between and depends on your overall playing style. But it's never 60% if we keep lucky streaks out of the discussion.

      /Johan = :f_confused:
    • Boomer2k10
      Boomer2k10
      Bronze
      Joined: 22.09.2010 Posts: 2,551
      Analysising W$SD without looking at WTSD is terrible especially over a small sample.

      If you're looking over 10's to 100's of thousands of hands then yes if you're W$SD is around 60 then you are either the luckiest person in the entire universe or you are showing down about 30% of the time and thus probably getting run over

      Why would you analyse just one stat and draw conclusions from it when you have all of them available to you? Use ALL the data available to get a conclusion otherwise your analysis is incomplete and your conculsions are most likely erroneus

      As per example from my last 1000 hands my W$SD is 55. What does that tell you about my play? Nothing is the answer.

      What about how my VPIP and PFR is 37/30 and my 3-bet is 17?

      Ok now we're getting a picture than I've probably been getting good hands preflop but I also could just be a massive psycho or sat in games where I have 2 empty chair style players (22/15 or something) to my left so I've got ballistic from the HJ, CO and Button. Or I could ahve had a lunatic to my right who plays awfully in 3-bet pots so I 3-bet with a massive range. Or have i just been playing in Super Shorthanded games which are nautrally more aggressive?? Since can't draw a good conclusion really.

      Ok now let's add the Fact my WTSD is 45....and now we have the conclusion. Boomer's had a horseshoe shoved up his ****.

      But there is no reason whatsoever to treat one stat as more important than another, certainly not a base stat, obviously things like "Bet when checked to in 4-way pot on River when it's Sunday" isn't important but a straight up main stat they're all important and all form a picture.
    • YohanN7
      YohanN7
      Bronze
      Joined: 15.06.2009 Posts: 4,086
      @Boomer

      The thread is about WTSD and W$SD. Nobody (at least not me) is saying "ignore the rest of the stats". I though that I made it clear in my last post that I didn't mean that one should check the opponents W$SD for the last ten hands or so. I was talking about [I]our own long term W$SD[/I]. If you have a W$SD of 60%, then it is a simple mathematical fact that you will improve your results by seeing a showdown more often when you have seen the river. It's less trivial to show that you fold too much on the turn, let alone on the flop. But it is clear that once the cards are turned over, you are crushing your opponents ranges. Your second paragraph seem to acknowledge this. Still, I get the impression that you crush my reasoning completely.

      But I stand by what I wrote. The W$SD is one of few stats that alone can tell you something about your game. That doesn't mean that you should treat it independently of the other stats. The 60% W$SD is probably only part of the problem.

      (I hate being told that I'm terribly wrong. That's one of my very many bad characteristics :rolleyes: However, tomorrow I'll probably admitt that I am bananas, but not just yet :D )

      /Johan :f_confused:
    • JLeitmotiv
      JLeitmotiv
      Bronze
      Joined: 02.01.2009 Posts: 756
      To explain this a little more theoretically, let's analize boomer's stats over those 1k hands. Assuming he saw 37% of flops (I don't think he doesn't get enough freeplay from the BB to change that stat), and his WTSD is 45%. That way, the 1k hands, only 0.37*0.45*1000=166 saw showdown. If we want to determine his W$SD from 55 with a 95% confidence interval, we would have something that has a range from (I found the wiki article here).


      W$SD: 48-62

      What I mean is that you need many many hands to make W$SD a significant stat (the less it happens, the harder it is to go to its statistical mean), and luck can (and will) modify a lot that value over different sessions.
    • YohanN7
      YohanN7
      Bronze
      Joined: 15.06.2009 Posts: 4,086
      Yup!

      Boomer is right too. I don't have W$SD on my hud for the reason that it is mostly useless for reading opponents.

      But you do have your own stats. That W$SD should be reliable if you have played enough.

      /Johan = :f_confused: