Is This Possible?

[Monteroy]

An odd discussion broke out at an Omaha table this morning when one of the other players made a comment about his pokertracker stats of a maniac at Absolute poker.

His claim was to have over 10,000 hands of the maniac who's only play is to go all in preflop at the no limit Omaha hi lo tables, while doing this every hand.

The claim is that this maniac is winning at around a 22 BB/100 hands rate.

I found this to be a bit unlikely without at least some elements of the story being off, but I admit I was not quite sure how to even calculate the odds of this taking place, and frankly a lot of the discussion collapsed into my trying to explain that assuming every hand is all in preflop it really does not matter who is playing the hands since a trained seal could be playing the all in every hand.

As best as I can tell, here are some of the variables.

- Person plays every hand all in preflop.

- Opponents choose when to be all in. I assume hand's played by opponents will be at least slightly better as a result (while the maniac in question has to play hands like 9222).

- No idea how many all ins go uncontested, but I assume it is not a huge number.

Overall win after 10,000 hands is around 2,200 (22 BB/100 *100) big blinds so using $25 buy in with 10 cent/25 cent blinds that would be around a $550 win.

I certainly believe players can win 22 BB/100 in Omaha hi lo by bring a lot better then others over 10,000 hands ( a point the person making the claim kept saying), but it seemed to me that since play skill was completely removed from the maniac's side it just came down to expected value, sample sizes and probabilities.

Any thoughts on the possibilities of the above being true are appreciated.

Thank you in advance.

[T50_Omaha8]

google:1 against

[Silent A]

Impossible, especially when you consider the fact that the rake will eat into this guy's profits like nuts.

22BB/100 is crazy enough for a zero skill strategy. When you add the rake he has to cover (which will be insanely high, like 4 BB/hand called or more).

[Silent A]

I decided to try to calculate the (im)probability of this working. First, some back ground assumptions:

1) Hero bets $25 every hand, 9 handed
2) 1/2 the time he wins the blinds
3) 1/2 the time he's called by one opponent
4) when called his equity will be: scoop 1/3 of the time, split 1/3, and lose 1/3.
5) Hero plays 10000 hands with a win rate of 22 Big Blinds/100 hands or more.

Notice that I'm actually giving him a +EV strategy before the rake, so I'm being very generous.

For the 5000 times the blinds fold, he wins $1555 (taking into account the times he's in the blinds).

When he splits he will lose an average of $1.15 due to rake.
When he scoops he will win an average of $22.70.
When he loses, he's out $25.

If we assume the splits come with the expected frequency (1334) he loses $1534 total.

His net win from stealing blinds and scoops is = 1555 - 1534 = $21

For a 22BB/100 win rate, he needs to win an additional $529 form scooping more than losing.

To do this he needs to win 1409 out of the remaining 2666 pots or more.

The probability of doing this is about 1/620 if he's 50/50 on each pot.

If we reduce him to something like 45/55 (which is still pretty generous) we get something on the order of 1 in 1,000,000,000,000,000.

If we put his equity any lower, Excel can no longer calculate a probability any higher than zero.

[Silent A]

[ QUOTE ]
google:1 against

[/ QUOTE ]

How you came up with this?

Interestingly, if you try to do this even more realistically ...

Say the table still folds half the time, but the other half he gets called by a one top 10% hand. A random hand vs. a top 10% hand will split 30%, scoop 25% and lose 45% for a total equity of 40%.

If you run these numbers using the other assumptions in my previous post, in order to show a 22BB/100 (big blinds) profit, he must scoop 1850 of the 3500 hands that don't split. In these scenarios, he is a 35/65 dog in each hand.

If you do the math, this requires him to run more than 21 standard deviations above his long term average.

And the probability of that is?

about 1 in 10^99.

OK, I have officially beaten this to death. I can now sleep.

[BradleyT]

You sure the guy goes all in every hand and not every playable hand???

[Bantam222]

Are you sure this isn't just a shortstacker and he folds all his junk and pushes a small range?