Please give your comments !!!!

@danvari. I can tell the difference between running badly and playing badly. I play badly now. Even when I play okay I am aware of mistakes worth 4BB/100 hands on average. (I do review my hands you know.) Add to that room for improvement worth of 2BB/100. This would make 8BB/100. It isn't easy, but in no way takes a world class player to do it. You can constantly find tables with dofuses capping all streets with trash, stations calling 90% all the way to showdowns, etc. Table select, play well and concentrate and you can do it. But right, this is probably impossible above the micros.

Disposing of 30k hands of bad results as probably variance and something that is supposed to happen is making it too easy for yourself, especially if you know (as I do now) that you don't play your A-game.

What mathematical model do you use when you calculate the chances of an Xk hands YBB/100 relative downswing compared to your standard ZBB/100 winrate?

Right, you can play a lot from the blinds in limped pots. Not so in raised and reraised pots. This is what I said. But I also said (or meant to say at least) that position works against you, something you seem to ignore. I am aware that posted blinds count towards entering the pot.

http://pokerdope.com/poker-variance-calculator/

Originally posted by YohanN7
Even when I play okay I am aware of mistakes worth 4BB/100 hands on average. (I do review my hands you know.) Add to that room for improvement worth of 2BB/100.

You probably think you could have improvement rates in that range. But the reality is this is far away from possible. Edges in fixed limit games are too thin. No winning FL-player can improve his winrate by 6BB/100. It's not even possible to have an overall winrate in that range. Ask any player playing for a few years and he will tell you exactly the same.

Of course you see lots of mistakes when reviewing your hands as you now have complete information.

Lets try some calculation:

I don't know what's your VPIP, but to make it easy lets assume it's 30%.

So you play 30 hands out of 100. As a winning player you should win a little over 40% which equals ~12 hands. How much do you think you can improve in those 12 hands? You just can't bring the other players to give you more money if you are holding the best hand. You can probably start value betting a little thinner or you could try to bring in some check raises oop. But if you are already a good player, there is not much room for improvement during those 12 hands. If you are trying very hard you can probably do something like 0.5 BB/100.

Now we are left with 18 hands. From those 18 hands you will have to c/f flop a little over 30% (~5.5 hands), because you didn't hit anything. There is absolutely no way to improve your winrate in those hands. The only thing you could try is to bluff people out of their winning hands (often multiple players). Try it on microstakes FL and you will see what happens.

Now you are left with 12.5 hands where you have to lose less or bluff more. The majority of those hands left you will play oop. To make your improvement rates possible you need to improve by 0.44BB per hand. In most of those hands you will either have a weak hand or a draw or a combination of both and you will not make your hand by the river. But you will often have to call flop and/or turn because of odds. Other times you will make a good hand and still lose.

How do you think you can have improvement rates in your given range? Please explain.

I will answer more later.

Originally posted by YohanN7
This would make 8BB/100. It isn't easy, but in no way takes a world class player to do it. You can constantly find tables with dofuses capping all streets with trash, stations calling 90% all the way to showdowns, etc. Table select, play well and concentrate and you can do it. But right, this is probably impossible above the micros.

I stick to what I said before. Not even the best player in the world would be able to do this on any stakes over 0.02/0.04. I don't even know about what stakes we are talking. You can probably tell me.

Originally posted by YohanN7
Disposing of 30k hands of bad results as probably variance and something that is supposed to happen is making it too easy for yourself, especially if you know (as I do now) that you don't play your A-game.

Just to clarify things, I should tell you that if 100k hands of FL is your lifetime sample size, you don't even have reached the 2% mark of my lifetime hands in this game. How likely do you think it is that you know more about long-term variance than me?

Bad runs can go over the 100k hand mark. Try a few million hands and we can talk again. On top of that it's just not possible to play your best poker when you take bad beat after bad beat. You can try as hard as you like, but if you run bad over a long time you will always play worse than during times when you get your fair share of pots. It's just human to do that. We are no robots.

But I also can understand why you believe you could do that. I probably believed something similar when I started playing. But if you keep playing the time will come when you will remember this. I'm 100% sure about that!

Originally posted by YohanN7
What mathematical model do you use when you calculate the chances of an Xk hands YBB/100 relative downswing compared to your standard ZBB/100 winrate?

The link Avataren posted uses the same formula as I used for the calculation. But it's not made for the calculations I did. It's not very flexible.

Would you like to know more? Confidence interval

I want some bonus points for quoting Starship Troopers!

You can use excel for your own calculations. As I use a german excel version it's most likely not very helpful if I tell you the formula in my excel is KONFIDENZ. I don't know what it is in other languages. Most likely the english version is CONFIDENCE.

If you put your winrate of 2.4BB/100 for a 100k sample size with a standard deviation of 20BB/100 (not sure if this is right for your playing style) into it you can calculate that your true winrate is between 0.77 and 4.03 with a probability of 99%.

The probability that your winrate is anywhere near 8BB/100 is roughly 0.000006% in a hypothetical model, which means transfered into the real world it's non-existent. I know you never mentioned that you can do that at this point. That's just for fun.

Originally posted by YohanN7
Right, you can play a lot from the blinds in limped pots. Not so in raised and reraised pots.

It doesn't count often for reraised pots, but for sure it counts for raised pots. Especially if more than one player already entered the pot, which is highly likely on microstakes. The price gets just too good to fold a lot of hands.

Originally posted by YohanN7
But I also said (or meant to say at least) that position works against you, something you seem to ignore.

In a game where most hands will go to showdown and the table is full of bad players position gets less important. It's much more valuable in very tight games or in No Limit - even more in PLO. But i think we are talking about very loose FL games.

If you think you are the best player on the table you should try to get into as many +EV situations as possible. The money is made postflop and players will make lots of mistakes. Combined with the good odds you get for playing a hand from the blinds there is often no justification to fold preflop.

In super loose games it should be possible to have a nearly breakeven winrate in the SB.

@danvari76
First off, when I review my hands I do not consider what the opponent shows up with. You probably don't do that either, so you shouldn't assume that I do it. That is, I can find mistakes in my play for a net worth of 4BB/100 hands, not result-oriented afterthoughts (oh, I should have raised, my 9 kicker beat his 8 kicker - damned).

You should be more careful about your usage of the word "impossible". A few years ago the best robot (then Polaris) was unbeatable by humans. But somebody associated to the research team claimed it was beatable for 20BB/100. Ridiculous - they said. It did turn out that it was beatable for 12BB/100 ( by a "nemesis" specifically designed to beat the shit out of Polaris). This principle is what I am referring to when I say that micro stakes players are beatable for a tonne.

If the "unbeatable" Polaris was beatable by 12BB/100, then what do you think you could cream out of a below standard micro stakes player?

That is, you first of all eradicate as many of your true mistakes as possible, i.e. such mistakes that count as mistakes in a thorough hand review using no assumptions about the opponent. I have at least 2-3BB/100 worth of those to take from. This means something like 10 mistakes per 100 hands worth on average 0.2BB each. This happens easily in multi-way pots. I play something like 32/23, add to that plenty of hands where I don't VPIP, but get a free flop in the BB. At a good table I VPIP 40% (because it's cheap to enter limped pots from the blinds). This means a shitload of decision points/100 hands = plenty of room for mistakes.

Then you should enter into maximum exploitative mode. That is, you become everybody's nemesis to the extent that you can. This is something that requires not only bad opponents (in the sense that they make bad moves that they think are correct (e.g limping on the button)), but opponents that don't notice what you do. Thus table select carefully! Now a second category of mistakes appears. These mistakes can be classified as moves you wouldn't make if you considered how this particular opponent is playing. For example, there are "regs" that I have 95% accurate reads on. Like if they raise flop, then they do not have a very good hand (because they always wait until turn to raise with a good hand). Failing to 3-bet a good hand here is probably a 0.3 BB mistake on average. Failing to snap-fold your big pair on the turn when the passive fish wakes up can be a close to 2BB mistake on average depending on the opponent. Likewise, failing to call down with your big pair (or even reraise) can be a 2BB+ mistake versus other opponents. Know your players. (I recall a recent hand evaluation thread where everybody (winning players!) advocated shoveling BB's into the pot with both hands. The judge (correctly) said passive calldown. Going berserk in big pots can be huge mistakes, not 0.1BB mistakes, versus some players.) This is where the big win rate can be found. You are a nemesis.

So, yes I do believe it is possible to beat the micros with 6-8 BB/100 ( if you allow me to develop a robot - remember Polaris - 20BB/100 - at least). If this wasn't the case, then it would be totally impossible to beat low-stakes for 3BB/100, let alone to beat high stakes for even 1BB/100. There would be no winners, only rake contributors.

But this is not really something I strive for. It makes multi-tabling impossible. It takes away some of the fun. I am not the kind of guy that enjoys single-tabling (or two-tabling) the micros to the extent that I closely watch every showdown and take meticulous notes on everything. It's just not worth it, but it is possible.

And, oh, you also shouldn't assume I began playing poker yesterday. Nor should you assume I'm totally lost at math (a hobby of mine).

It is necessary to add here that mistakes according to exploitative play need not be mistakes in the sense of GTO. Making a tight fold in a 10BB pot that turns out to be a 0.05BB mistake (or even correct) under GTO may be a 3BB mistake on average given that we are up against a maniac. Do these situations occur at the micros? Yes, they do!. You should look for them. With two maniacs at the table, there is probably room for plenty of ordinarily small mistakes that get magnified totally out of proportion, both folding and calling incorrectly can be extremely costly.