PDA

View Full Version : Computer simulation problem

microbet
10-21-2005, 12:27 AM
Ok, this belongs in theory or some forum that I never post in, or maybe MTT and maybe I'll cross post, but I don't know anyone there and I really want the best shot at a particular member to respond. (you get three guesses and the first two don't count)

You have 5000 players and they come in two flavors. Whenever a FlavorA player plays a FlavorB player the FlavorA player is 60% likely to win a small amount of chips and 40% likely to lose a small amount of chips. When a FlavorA player plays a FlavorA player they are 50/50 to win/lose a small amount of chips. When a FlavorB player plays another FlavorB player they are 50/50 to win/lose a large amount of chips.

Construct a tournament where there are only HU confrontations, obviously, and play with the ratio of FlavorAs and FlavorBs and also what a small number and what a large number of chips means and see what happens, perhaps over a large number of tournaments.

Feel free to either program the simulation or just go stream of consciousness on it.
Exitonly
10-21-2005, 12:52 AM
i'm not sure what you want here.. 'FlavaA's are obviously going to be +EV in all situations that have atleast 1 FlavaB. Higher EV for deeper structures, or more FlavaB's.

maybe a bit more info? or i might just be too tired to answer this.
eastbay
10-21-2005, 12:59 AM
[ QUOTE ]

You have 5000 players and they come in two flavors. Whenever a FlavorA player plays a FlavorB player the FlavorA player is 60% likely to win a small amount of chips and 40% likely to lose a small amount of chips. When a FlavorA player plays a FlavorA player they are 50/50 to win/lose a small amount of chips. When a FlavorB player plays another FlavorB player they are 50/50 to win/lose a large amount of chips.


[/ QUOTE ]

I like it, just not enough to be pulled off my current project.

I think you're talking about Slim Pickens, by the way. I bet he can whip this out in under an hour. /images/graemlins/wink.gif

eastbay
Big Limpin'
10-21-2005, 01:26 AM
Flava Flav crushes them all, end of story.
http://www.unc.edu/~hymas/images/flava_flav1.jpg
YEAHHHHH BOOOOIIIIEEEEE!!!!
microbet
10-21-2005, 01:43 AM
I'm thinking you have a lot of FlavorBs compared to FlavorAs.

What I'm really wondering is, in a big MTT will the slow and steady player ever accumulate enough chips to be in a good position against the luckiest of the luckboxes.
Deuce2High
10-21-2005, 01:50 AM
What I'm wondering is how can you guys even start guessing what the outcome is when you have so many variables yet to be set? It looks like a great experiment to try out, but without setting up a program and entering variables it seems like a waste of time to guess what's going to happen.
microbet
10-21-2005, 02:01 AM
Who guessed what will happen?

Ok, I'll guess. I think FlavorAs will win more than their share almost regardless of the ratio of types of players or sizes of wins and losses.

Ok Slim, warm up the vacuum tubes, or grease up your abacus or whatever you need to do to program this in your wierd language.
Exitonly
10-21-2005, 02:25 AM
it's not really a guess... over long term FlavorB will have neutrral EV at best, and if there are ANY flavor A's it'll be negative.

opposite for the A's, at worst Neutral, +EV more and more for more B's. Or deeper structure.
microbet
10-21-2005, 02:35 AM
Well, that's my thought too, but I'm not sure about a situation like this:

You have 4000 FlavorBs and 1000 FlavorAs. After a while, say 3500 FlavorBs are busted or close, but the 500 that aren't have HUGE stacks and the 50 luckiest FlavorBs have HUGE HUGE HUGE stacks and if you add something now like escalating blinds the FlavorAs will never catch up.
10-21-2005, 02:53 AM
[ QUOTE ]
Well, that's my thought too, but I'm not sure about a situation like this:

You have 4000 FlavorBs and 1000 FlavorAs. After a while, say 3500 FlavorBs are busted or close, but the 500 that aren't have HUGE stacks and the 50 luckiest FlavorBs have HUGE HUGE HUGE stacks and if you add something now like escalating blinds the FlavorAs will never catch up.

[/ QUOTE ]

Right, if the FlavorAs are sufficiently outnumbered, it's unlikely one of them will come in first, but they will still have greater EV per player.

Am I missing the point?
Slim Pickens
10-21-2005, 04:53 AM
So here's what I did in one hour:

2 Type A's and 8 Type B's

A's vs. A's: The amount wagered is 1/4 of the smaller stack, unless the difference between the two is greater than a factor of ten, in which case the wager is the entire amount of the smallest stack.

A's vs. B's: Same as above, except A's win 60% of the time.

B's vs. B's: The wager is the smaller of the two stacks.

They start with 20 chips each and go until someone has all 200.
Player Type A wins 256/1000
Player Type B wins 744/1000
Slim Pickens
10-21-2005, 05:41 AM
[ QUOTE ]
I'm thinking you have a lot of FlavorBs compared to FlavorAs.

What I'm really wondering is, in a big MTT will the slow and steady player ever accumulate enough chips to be in a good position against the luckiest of the luckboxes.

[/ QUOTE ]

5000 players enter, only 100 are the world-class "A Players" as you defined, using the same HU confrontation rules I laid out... What I've made is essentially a "blindless" tournament, but with constant wager sizes normalized to the stacks left in play.

I played this 100 times and an A Player won it 61 times. That's pretty good for starting with 2% of the field.