Sit N Go Analysis

[imcastleman]

I’m developing a program that reads through a given set of my tournament hand histories and writes out an excel spreadsheet where each line item contains details surrounding hands where I am all in. I’m doing this so that I can take a particular days work and find out a few things. One, to find out how often I’m going all in as the underdog and two, to find out if I was lucky or unlucky on certain types of all ins(i.e./ coinflips, 60/40s, dominating hands and over-dominating hands) that day. I’ve got most of it complete but I’m having a difficult time with the probabilities for a given situation. What I’m hoping someone can do for me is the following:

If I give you 2 hands in the format just like www.twodimes.net/poker (http://www.twodimes.net/poker) and say hand one is in cell A1 and hand 2 is in cell A2. I’d like there to be a formula in cell A3 that calculates the probability that the hand from cell A1 wins. If this cannot be done easily, then for each day that I analyze, I will just manually go into the file and place the approximate probabilities for each situation, although I’d really like it to be dynamic if possible.

For those interested in seeing a sample spreadsheet, PM me your email address, and I’ll send a copy.

[Scuba Chuck]

IM, how do you see this benefiting you?

[Slim Pickens]

My first guess is that a 169x169 worksheet lookup table of equity would be the best. Can Poker Stove or similar generate this for you?

SlimP

[joeboe2001]

If you are going to this length, is 169 x 169 adequate?

I would think 2652 x 2450 would be necessary? (52x51 times 50x49) That would enable the program to take into account all possible hands, including suit interdependencies (i.e., if you have two clubs, you have a lower probability of getting a flush if your opponent also has two clubs?)

[maddog2030]

Open source (http://sourceforge.net/projects/pokersource/) is your friend. Pokersource does all the nitty gritty stuff so you don't have to waste time programming card calculations. And it's pretty efficient too. You can even enter hand ranges and such, etc.

If you're ever working on any interesting programs, drop me a line. Lately I've started writing a few to analyze certain scenarios I've come across, but have a few ideas for more semigeneral things I can do that would be more interesting. I'm always interested in hearing what others are working on.

[Slim Pickens]

Eh, yeah it's necessary to get the exact answer, but the computing cost might be high. I think the idea is just to get close here, with a few percentage points one way or the other not making a huge difference in the interpretation. Maybe I just think that way because I have a slow computer.

It looks like I have to "pay" for a "licenced copy of Office XP" to "legally use" the "features" necessary to retrieve data from a webpage.

SlimP

[Freudian]

[ QUOTE ]
IM, how do you see this benefiting you?

[/ QUOTE ]

It would bring bring a new era of bad beat posts.

"Damn, I have -7.4% luck in my last 100 SnGs"

[Slim Pickens]

Quote from eastbay in this thread. (http://forumserver.twoplustwo.com/showthreaded.php?Cat=&Number=2359781&page=&view=&s b=5&o=&vc=1)
[ QUOTE ]

Already did something like this awhile back. I called it "EV vs AV" (expected value vs. actual value).

I used it briefly but found soon that it was kind of a super techno pacifier that I would suck on when taking a lot of bad beats. "Oh yeah, look, I am running bad. But I already knew that."

[/ QUOTE ]

I thought this sounded familiar.

[maddog2030]

Yeah, there's been talks of "luck-o-meter's" before also. Maybe not the most useful thing in the world, but it would be nice to know if you're running hot or cold on your matchups. Maybe you know you're running cold, but sometimes you may think your upswing is based on your skill, but the luck-o-meter will let you know this is above and beyond your skill and into good luck territory.

Of course there's a lot more "luck" than that. There's a lot of components that you can't control that could all fall under "luck", like if you run into AA on the bubble every game for the rest of your life, etc.

[hesbala]

Do you have eastbay's program? There is a 169x169 EV chart (in notepad) distributed with the program.

[Slim Pickens]

Well there you have it... So it sounds like were within inches of a Luck-o-Meter. Let the Bad Luck posts begin.

[imcastleman]

This probably won't benefit me at all. I just have a job that doesn't allow me to play poker at work so I have to find something interesting to do with my little bit of spare time I have at work. In addition, I think it would be neat to see over the long run how many times I have the worst hand going in for every block of 60 tournaments. I may notice a pattern that on the days that I have a high roi that I was rarely going all in with the worst of it. I don't like going by feel because it always feels worse than it really is.


Nothing special in this analysis, but maybe I'll find something out about my game that I won't find otherwise. Plus, I have a coach who I'd like to be able to give my results to in a nice concise fashion(rather than 60 hand histories a day) and let him interpret what I am doing, whether it be right or wrong.

[astarck]

[ QUOTE ]
Yeah, there's been talks of "luck-o-meter's" before also. Maybe not the most useful thing in the world, but it would be nice to know if you're running hot or cold on your matchups.


Of course there's a lot more "luck" than that. There's a lot of components that you can't control that could all fall under "luck", like if you run into AA on the bubble every game for the rest of your life, etc.

[/ QUOTE ]

This is why I think a luck-o-meter will never be very useful. When you are running hot or cold it has to do with A LOT more than just 1 coinflip that you may have won/lost. The luck (or variance) of the cards dealt throughout the course of the tournament comes into play.

Obviously winning coinflips is all "luck" in the short-term and when you are winning those you seem to be on hot streaks. But getting dealt AA-QQ 10x in one tourney would be "luck" but wouldn't factor into the program. (Unless of course the luck-o-meter actually analyzed what cards you were dealt throughout the sng and checked for an average distribution. Or as even as you can get through 40-60 hands.)