some preflop all-in empirical data

[eastbay]

In my last 88 $55 PP games, I've gotten it all-in preflop 279 times.

It was folded to me 134 times, and called 145 times (!).

When called, my total expected value was -7k chips. (My actual value is -15k chips, but that's not the point here.)

My average expected value per hand when called is -46 chips per hand.

When folded the blinds, I pick up on average 384 chips. I picked up a total of 51.4k chips from folded blinds.

Discuss.

eastbay

[AleoMagus]

To me, this suggests that pushing all-in with a few more hands might be a good thing. You are getting similar call/fold figures, but a way bigger average profit/All-in than loss/call. This might mean even looser all-ins could be effective.

Hard to say though, as you'd obviously get more calls which would be beating you, but still, as is, your steals are earning you so much so I'd think a few more all-ins could not hurt as much as it would improve your results.

Just my initial reaction. Interesting stats without giving any other details about all-in hands, etc...

Regards
Brad S

[stupidsucker]

wow those are great numbers. Looking at them I agree with AM. You can probably push a little more and pick up a few more blinds.

Your EV when called is half as low as your AV, so your chip loss per hand should be less then it is too.

I should take a look at mine as well.

[rachelwxm]

[ QUOTE ]

My average expected value per hand when called is -46 chips per hand.


[/ QUOTE ]

This is interesting! Assume you have 800 chips on average when pushing, that suggests you are not that much behind when called. Either your opponent does not adjust the GAP well or you are too tight.
It will be more clear if you break down those by stages.

[RacersEdge]

If I read this right, another metric you can extract is you are getting 1.52 blind steals per game.

My comment on raising more is that there could be an interactive effect: the effectiveness of all-ins could drop if you become looser with them - but it still could be optimal to do.

[AleoMagus]

[ QUOTE ]
another metric you can extract is you are getting 1.52 blind steals per game.


[/ QUOTE ]

I would hope he's getting more than that. These are just the all-in steals.

Regards
Brad S

Edit: On average, I'd guess that successful players are getting 3-5 steals per SNG. That's just a blind guess though, so maybe it is closer to 3 than 5. On average in each SNG that I WIN , I'd bet I'm getting as much as ten or more.

[The Yugoslavian]

Eastbay --

First, this is a great post and interesting way to look at improving one's game, kudos to you.

I have several questions that would love answers to, but of course they may be too time consuming to have tracked:

Are those 279 times you went allin preflop the totaly number of times you were allin preflop or the number of times you went allin in an unraised pot or limped pot?

I would also be curious as to how many of those times you were on the bubble (4 to maybe 6 handed), early in the tournament, 3-handed and heads up.

Does the 279 number include allins from other players who had you covered and that you called?

What is your average position when going allin and how does it break down?

On average, what has been the size of your stack?

Also, did you track this yourself or is there some software that does it (or that you created)?



PS I'm originally from the eastbay so I'd have to say that I'm a big fan of yours, [img]/images/graemlins/wink.gif[/img].

[eastbay]

[ QUOTE ]
Eastbay --

First, this is a great post and interesting way to look at improving one's game, kudos to you.

I have several questions that would love answers to, but of course they may be too time consuming to have tracked:


[/ QUOTE ]

Good questions. I was being pretty sloppy here with my description of what I was reporting. Let me clear things up a little.

[ QUOTE ]

Are those 279 times you went allin preflop the totaly number of times you were allin preflop or the number of times you went allin in an unraised pot or limped pot?


[/ QUOTE ]

Actually it is the total number of times I was all-in preflop, period.

"Was called" really meant there was a showdown, no matter how it occurred. So if I called with all my chips, I misleadingly reported that as "was called."

[ QUOTE ]

I would also be curious as to how many of those times you were on the bubble (4 to maybe 6 handed), early in the tournament, 3-handed and heads up.

Does the 279 number include allins from other players who had you covered and that you called?

What is your average position when going allin and how does it break down?

On average, what has been the size of your stack?

Also, did you track this yourself or is there some software that does it (or that you created)?

PS I'm originally from the eastbay so I'd have to say that I'm a big fan of yours, [img]/images/graemlins/wink.gif[/img].

[/ QUOTE ]

All things I thought about calculating, but didn't make the effort yet.

I wrote a small script (program) that looks through hand histories to pull out this information. Adding those additional stats is do-able but I'd have to be convinced it was worth the time and effort.

I'll let you know if I refine the script.

eastbay

[eastbay]

[ QUOTE ]
To me, this suggests that pushing all-in with a few more hands might be a good thing.

[/ QUOTE ]

That was my initial reaction as well. And it makes sense, since in my last 100 games I had been experimenting with a fairly tight stealing strategy.

I think maybe for the next 100 games I will experiment with a fairly loose stealing strategy, and then run the same analysis (although doing it a bit more carefully in both cases - see my response below.)

eastbay

[eastbay]

[ QUOTE ]
[ QUOTE ]
another metric you can extract is you are getting 1.52 blind steals per game.


[/ QUOTE ]

I would hope he's getting more than that. These are just the all-in steals.


[/ QUOTE ]

Yep. I get far more than that if I place.

eastbay