Top 5% of NLHE hands, top 10%, top 15%, etc.?

    • FORECONSniper
      FORECONSniper
      Basic
      Joined: 27.10.2011 Posts: 2
      Hey everyone!

      I'm currently working on using SNGWiz to brush up on my optimal short stack strategy. And, as such, I have been working on putting villains on a preflop range of hands - currently trying to incorporate Pokertracker 3 HUD stats to formulate an accurate range of my opponents for STT SNG tourneys.

      My question is this: I use pokerstove to, for example, find the top 10% of holdem hands. I also used SNGWiz to determine the top 10%. Now, the problem here inlies: The "top 10%" of hands is DIFFERENT for Pokerstove as it is for SNGWiz. Now, I also understand that there have been a few webpages out there that list ALL NLHE hands by "top 5%, top 10%, top 15%, etc". And, again, as you imagined, each of these websites I've reviewed shows a different range for each of these categories.

      I would like to see if you guys out there have input as to what the MOST accurate chart for these are.

      I am working on using stats in PT3 HUD like VP$IP/PFR/Af to put these opponents on accurate ranges and therefore, need to know if you guys have a good recommendation as to what I should use to base my ranges on?

      I imagine that every villain has a slightly different way of categorizing these percentage hands so there must be a lot of variables that either make one's reads fairly accurate or fairly inaccurate.

      I also vaguely remember in one of my Sklansky books (probaby NLHE: Theory and Practice) that may have had a chart of all nlhe hands by order of top whatever %s.

      Please advise,

      FORECONSniper
  • 3 replies
    • pzhon
      pzhon
      Bronze
      Joined: 17.06.2010 Posts: 1,151
      Different hands are better for different situations, so there shouldn't be one most accurate list.

      Consider the Nash equilibrium calling range at 10 bb. It's about 38.5%: 22+ Ax+ K2s+ K6o+ Q6s+ Q8o+ J8s+ J9o+ T9s

      If you play push-fold at 20 bb, the Nash pushing range is about 38.8%: 22+ Ax+ K3s+ K9o+ Q5s+ QTo+ J7s+ J9o+ T6s+ T9o 96s+ 86s+ 75s+ 65s 54s

      Hands like K2s, K6o-K8o, Q8o-Q9o were part of the top 38.5% calling hands, but not part of the top 38.8% pushing hands, which included J7s, T9o, 96s+, and some lower suited connectors down to 54s instead. This should make sense. 54s is almost sure to be behind, but not by too much, which makes it a good semibluff when you push. You don't get good enough odds to call with it for 10 bb. A hand like K6o is often ahead of a hand in a wide range, but it doesn't have much equity when behind. You can call with it hoping that you are ahead, but it's not a great semibluffing hand.

      There are more subtleties. Even when you have a 10% pushing range, the best 10% may depend on whether you will be called by a tighter range or by a looser range. If you have a 10% range because you need a lot of equity but you will be called widely, then you might favor hands like ATo which have a lot of equity against wide ranges. If you have a 10% range because you are pushing through several people who each have a chance to pick up AKo or AQs, then you need hands which do better against tight ranges, and then ATo does not do as well.

      SNG Wizard uses the Sklansky-Chubukov hand rankings when you set a range for an opponent. This is better for calling ranges than for pushing ranges. It tends to overvalue hands which would be good in 2-card poker compared with suited connectors for pushing hands.

      I think it is far more important to estimate the range correctly for your opponents in % (and to take out the top few hands when your opponent would not make a play with a really strong hand) than to guess and adjust for which hands each player thinks are in the top 10%. In general, if you think your opponent is not playing the best x% of hands, this usually means you can play slightly more loosely, since the inconsistencies of your opponents tend to help your marginal hands. It is rare for this to be a large effect.
    • FORECONSniper
      FORECONSniper
      Basic
      Joined: 27.10.2011 Posts: 2
      Pzhon,

      Wow that was an awesome, insightful reply from you. Thanks!

      A couple of things I need to look into: The Nash equilibrium calling range you mentioned. I may have come across it in my studies but I doubt it as I would have remembered it - I have primarily been putting in time studying the books including Sklansky, Harrington, Negreanu, Doyle, Poker Mindset, Hellmuth, etc. and, in addition, been putting in steady hours online at STT SNGs (I started as a live game player). A lot of these key concepts including ICM play, $equity, calling/pushing ranges the books simply did not cover! It's amazing how much studying one thing leads you to something else entirely different! That's what I love about the game!

      Anyway, back on track here. What you have mentioned in regards to top pushing vs top calling percentage ranges completely make sense.

      So how would I go about familiarizing myself with stuff like you mentioned: "A hand like K6o is often ahead of a hand in a wide range, but it doesn't have much equity when behind. You can call with it hoping that you are ahead, but it's not a great semibluffing hand." <---------- that whole statement you made there makes sense to me when I read it, however, if you were the one to have asked me to explain the equity of a hand like K6o, I would not have been able to come up with the words! How do I get to think like that?

      And the part where you mention:

      "I think it is far more important to estimate the range correctly for your opponents in % (and to take out the top few hands when your opponent would not make a play with a really strong hand) than to guess and adjust for which hands each player thinks are in the top 10%."

      that brings up another, more specific question:

      in trying to analyse correct/incorrect decisions in push/fold situations I have made using the PT3/SNGWiz combo, I came up against this issue: I loaded the same hand into both programs in a situation sort of like this:

      I was at a 10 man STT SNG with 5 left, top three pays 50/30/20%, blinds 200/400:

      Hero (BTN): 1670
      SB: 1780
      BB: 5120

      Folded around to me on button. I push with K2s. SB pushed, BB folded. SB called with A9s and won with flush by river. Now, at the time I made the play, I felt that this was sort of a no brainer. 4.5x big blinds left in my stack, looking at just under 3 "M" - red zone. Based on observation so far in the game, now in the late stages, I figure both blinds behind to probably call relatively tight (we have obviously said ranges are subjective earlier, but let's say for clarity's sake: the sb with a calling % of 15% and bb maybe 20%) figuring based on $equity, it wouldn't be "smart" for either to call much looser.

      Sure enough, the sb wakes up with A9s, and why wouldn't he call? He obviously pushes and wins as I stated above.

      When I run this hand in SNGWiz, the default calling % #s are: sb 29%+ and 37%+ for the BB. First off, how did the program arrive at these defaults and are they accurate to use? And under what conditions are they accurate? The Wiz suggested Fold.

      When I review the same hand in Pokertracker, using the Instant Hand Reply mode with HUD stats enabled, I assume (again, pretty new at using both softwares) that checking under the Preflop stats and looking for "Call PFR" would be the best stat to manually override the SNGWIz default ranges with? It was between that and CCPR (I think it is that... the Cold Call Preflop raise stat). Now, I only took the Call PFR stat and plugged it into Wiz adjusting from the defaults of 29%/37% to now my PT3 numbers of 20%/ 13% and now SNGWiz suggests Push.

      Tweaking these numbers obviously made Wiz suggest a Fold to a Push. This is interesting as you had just mentioned to me a hand sort of similar to this one. So, how do I find out what the most correct play in this spot would've been? Am I even examining the correct stats withing Pokertracker to do this with?

      Any other suggestions?

      Thanks for your help!
    • pzhon
      pzhon
      Bronze
      Joined: 17.06.2010 Posts: 1,151
      So how would I go about familiarizing myself with stuff like you mentioned: "A hand like K6o is often ahead of a hand in a wide range, but it doesn't have much equity when behind. You can call with it hoping that you are ahead, but it's not a great semibluffing hand." <---------- that whole statement you made there makes sense to me when I read it, however, if you were the one to have asked me to explain the equity of a hand like K6o, I would not have been able to come up with the words! How do I get to think like that?
      I have spent some time with PokerStove looking at the hot-and-cold equities of hands versus each other and against ranges. I think many poker players know match-ups like AK vs AQ, AK vs 22, and AA vs. KK, but far fewer know A4 vs A2, A3 vs 76s, QQ vs. 76s, or T8 vs 96. When you understand the individual match-ups, then equities versus ranges make more sense.

      Some more information about hot-and-cold equities went into the book I wrote with Collin Moshman, The Math of Hold'em.

      When I run this hand in SNGWiz, the default calling % #s are: sb 29%+ and 37%+ for the BB. First off, how did the program arrive at these defaults and are they accurate to use? And under what conditions are they accurate? The Wiz suggested Fold.
      The default ranges in SNG Wizard are unreliable. They are often wrong and need to be adjusted. SNG Wizard has a very simple model built in which does not adequately account for many important factors such as position. So, it assumes that the BB will call surprisingly widely against some early position pushes but almost the same range against small blind pushes.

      You can't simply use your HUD statistic for the call % of the blinds. This statistic lumps together many situations. Your HUD will show the same percentage whether you are pushing 4 bb or 12 bb, and yet the blinds will call with different ranges. You can try to stereotype players, e.g., try to determine what a 50/4/0.6 player will have when calling a 5 bb push near the bubble. You can use your HUD stats to identify which stereotype to apply.