PDA

View Full Version : Need ideas for Gambling related bachelor project is statistics.


Ratamahatta
10-18-2006, 08:41 AM
I study statistics/probability theory in university of Copenhagen and I have to write a bachelor project. I have to start after christmas and I will have 5-6 month for it. BP in our university works something like this: if you don't have any ideas, then you will get a data set (usualy data from clinical trial) and you have to do usual statistics (building models, making tests etc) using the data set.

Since I like poker and gambling (maybe I will end up working in gambling industry, who knows) I would like to write a gambling related project. I have been thinking about what I should write about but I couldn't come up with a subject. I have only one idea: I was thinking about finding a rigged online casino, deposit some money and play their rigged black jack (and maybe some other games) while writing down the result of every hand. Then I could make tests and "prove" that the casino is rigged (maybe even finding an algoritm that makes their games rigged). I'm not sure that I can write 80-100 pages (I belive this is how much I have to write) about it and if it will be sufficient enough to pass.

Does any of you have any ideas? Any advice/idea will be highly appreciated. Thank you!

Darryl_P
10-18-2006, 09:03 AM
A couple of ideas:

1) Examine a large number of online poker tournaments and observe the distribution of results on the next tournament played by the winner of the previous one. This has implications on the amount of skill there is in poker vs. luck.

2) Study the predictive power of late line changes in sports betting, say. Test a betting strategy of betting on any team whose line moves at least x points in the last y minutes before the game and see if the results differ from what you'd expect from chance. If this shows something significant, you might want to keep it a secret and use it to make a living.

3) Study the difference between home advantage of teams with high average home attendance vs teams with low average home attendance to isolate how much the number of attendees affects the outcome of the game. After all, home advantage comes from many other things as well such as being better rested from not travelling, playing with familiar surroundings, etc.

Just a few off the top of my head...your test for cheating sounds like a good one too. And if you can't fill 80 pages, you might try more than just blackjack...

Utah
10-18-2006, 09:06 AM
Here is one that quickly popped into my head and that I have already done to some degree. I am not sure if it fits what you want to do:

What is the probability of a players preflop holdings given: His playing history, his action, his situation (position previous raises, previous calls, position, etc), the history of the cards he has shown down in the past, the percent times he goes to the river in certain situations, how similar players play, the odds of being dealt certain holdings, etc. And, what are the odds of certain hands making it to the river in certain situation.

Basically, you want to makes guesses based on samples and incomplete info. This is a very hard problem but it could be greatly simplified for a class project.

Anyway, just one idea. Feel free to discard.

Note - if you do a poker problem I have billions of real money limit game records and could get you all the data you need.

FortunaMaximus
10-18-2006, 09:08 AM
/images/graemlins/frown.gif Any finite system is rigged. /images/graemlins/tongue.gif

Of course there's a flow and nuance to a large field of patterns. That's unavoidable.

Can someone really write 80-100 pages about it? Wow. Seems like such a waste of trees.

FortunaMaximus
10-18-2006, 10:14 AM
[ QUOTE ]
Note - if you do a poker problem I have billions of real money limit game records and could get you all the data you need.

[/ QUOTE ]

This intrigues me. Road trip for a couple days, but I might have an idea or two for a sample size that big. Hmm.

Ratamahatta
10-18-2006, 02:15 PM
Thanks a lot. I like Darryl_P's 2nd and 3rd idea. I don't know much about sports betting but I think that I will educate myself on the subject because it looks like there is a lot of sports betting related topics I can write about.

Utah's suggestion is good too. If I will write a project about it, I will then make a poker bot! /images/graemlins/tongue.gif

If anyone have other ideas, please post them, anything will do! I would like to see more sports betting and casino games related ideas as well.

Thank you for your help.

FortunaMaximus
10-18-2006, 05:43 PM
[ QUOTE ]
[ QUOTE ]
Note - if you do a poker problem I have billions of real money limit game records and could get you all the data you need.

[/ QUOTE ]

This intrigues me. Road trip for a couple days, but I might have an idea or two for a sample size that big. Hmm.

[/ QUOTE ]

I'd have to look at initial starting conditions to generate the picture, but I can't help but think it'd make an interesting fractal topography model.

That's the first idea. Hmm.

Sorry for the hijack, OP. GL with your project.

Utah
10-18-2006, 07:17 PM
[ QUOTE ]
I'd have to look at initial starting conditions to generate the picture, but I can't help but think it'd make an interesting fractal topography model

[/ QUOTE ]
Would this be to solve a real world problem or to make a really cool fractal picture?