KQ value?

[Peter-23]

All though this may seem to be a complaint I really have serious intent.

Over a sample size of about 35k hands on party.

I have had KQ (suited or not) 408 times.
I have had KJ (suited or not) 423 times.

Im ahead on KJ with 70BB
and behind on KQ with 50BB

Obviously KQ is a stronger hand so if the sample size would be much higher things would be reversed.

Now to my observation.

Always when someone place a question like this everyone cry out "blip sample size". This may be right but how big a sample size do I need to say anything about my game?

I mean if I need say 100k hands of say KQ to say anything about my game on that hand, then that must be true for any hand. If it’s true for any hand then I will need an overall sample size of millions of hands. Even when multi-tabling that will require multiple life times!

So, since it’s not realistically possible to accumulate enough hands then hand level analysis must be useless, right? (Actually played hands I mean).

My conclusion is that if all above is true then hand analysis on hand level can only be done theoretically! Making player related hand analysis virtually useless!


What do you say?

[disjunction]

One thing to note is that when you look at your own personal statistics, this is a lot different than, someone on here reading the statistics of some random person on a message board. When a weird stat is posted, everyone reading the message board has to divide its probability by 1000, because there are 1000 people who would have made the post if they encountered the weird stat.

I use weird stats as an early warning system. It tells me what to look at. I look for a common-sense explanation that would represent a flaw in my game.

The specific example you provided of KQ vs KJ doesn't seem too weird to me. 20 BB is only a couple of pots. If it were more than 20BB, the common-sense thing to look at would be how you play 2 overcards.

[Peter-23]

I think you misread the difference. I’m ahead 70BB on KJ and behind 50BB on KQ.

The difference is 120BB.

I play 1/2 so it’s much more than a couple of pots.


But, my point is that since even 400+ samples of data is not enough to say much then the “warning signal” might be misplaced (not necessarily though).

To go over how you play overcards is hardly a mistake in any event but if you were to change how you play. For example not playing KQ in early position (if you found you lost most from there) then you might actually give money away!!!

If my observation is correct then if theory says you should play KQ in early position you should never change that game plan, even if you have 400+ samples that says otherwise. Actually you could never say anything based on a particular hand stats from your own play because humans never live long enough to have required sample sizes!


I am still not sure my reasoning are correct but I start to believe that placing to much emphasis on your own game might be misleading. I think if you have a problem with your game you should check against theory and have theory as guideline. If you find you play in accordance with theory and still are behind, regardless of sample size, then everything is just fine. You will pick up or the theory is wrong.

Actually I think this is more interesting than first meet the eye. This suggests that there is a mathematically perfect way to play any poker game. The skills in poker would then be based solely on a player’s ability to find that perfect way based on specific game conditions in a specific hand against specific players.

I’m I taking this too far?

[sMethod]

NO. Some times pokers players can lie awake thinking of the perfect way to play the game. If you were going to play perfect you would know everybodys cards and the cards that were coming and how every player would act to ever round of betting to you could always maximize your wins and minimize your loses to an extent where you were unbeatable in almost every session. This is not reality, In real life we never know what card is going to be peeled off next and we dont know how any body is going to react, there are end less numbers of variables that factor into humans making decisions so it is impossible to predict there actions so we can never play perfect. Iv gotten off topic. Refocusing on your KJ vs KQ. 120BBs is a big swing but part of the natural swings that occur within the game. There are storys of pro players that have logged a losing year. Its not uncommon. Thats why you must be so phychologicaly strong to play this game it will eat you up if you let a bad run get to you.

I can see that your a very analytical person as am I but some times we have to trust theory and put our faith in the math of the game and just play as many hands as we can. /images/graemlins/diamond.gif

[FiReSiStAnT]

I like this point and this post

It is obvious that studying reasonable amounts of hands is very important to improve your game. It is important to watch your stats and how they change and realize why.

F*ck the Sample Size

[IcarusFalling]

i dont understand how it is possible for a pro to have a losing year unless there are other factors .. for example a BIG loss at a high limit game that crushes all their profits at a lower limit.... or only playing against people as good as you..

[IcarusFalling]

KJs and KJo are very different hands.. as well as KQs and KQo so i would not consider them as one... besides that it is interesting that you brought this up.... having KQo 400 times alone i would think should show a profit.. you'll hit your pair 1/3 times.. but it depends on a lot of things... do you play it against a raise when everyone else foldeD? do you play it 1 on 1? .. how tight are the tables that you play? i believe little things like this can make a dramatic effect and create more variance.. where as if 4 people to the flop (plus blind) every time im sure you'll be making some $$$ ~ ~ ... analyitcal analysis of many different things can create a larger understanding why certain hands have the statistics they do realistically if you get KQs 400 times you'll really only make the flush draw 50 times .. and the flush about 16 times.. and ask yourself how many of those times is there a possibility that you got a bad beat? ... im looking at my statistics and for me AJ is giving me problems.. now i find that AJo with a sample size of 80 hands has won me .14 dollars /images/graemlins/laugh.gif not much but it was a loser before this... now my sample size of my AJs is only 27 and it has lost 30BB (30 dollars).. now this is a much smaller sample but it might give you a better understanding (btw i have changed how i play AJ slightly and do believe it is helping)

[PairTheBoard]

How much differently do you play KJ compared to KQ?

PairTheBoard

[Peter-23]

If my thinking is correct I would argue that changing how you play AJ might be a mistake.

1. If the reason for changing is that AJ do bad according to you stats you make a mistake.
2. If the reason is strategic or theoretical it’s not a mistake.

For example early in a tournament then AJ, KJ and many more hands should not be played. Why because the risk vs. return does not warrant play. You may of cause argue that this strategy is wrong and perhaps you are right if you do. But the important thing is that it’s strategic and you change your play on those grounds.

Another example, say someone do not know how to play flush-draws. Then at least marginal suited hands should be folded pre-flop. This is of cause a simplistic example but it describes my point. Why put chips in there when you don’t know what to do.


The reason that 1 is a mistake is that you don’t know if you are doing bad since the required sample size takes more than a lifetime to accomplish.

Let’s take the classic dice example.

You get $6 if you hit a 6 and it cost you $1 to play.
If you play 50 times and hit a six only three times should you stop play just because you are running badly?