Question About Winrates

[JasonP530]

What is a good ROI for the 50+5 Single Table tournaments? I have been doing well, and wanted to know if my winrate is sustainable. How many tournaments do you need before you start approaching something of significance. I assume ROI includes the vig.

thanks for your help.

Jason

[The Yugoslavian]

[ QUOTE ]
What is a good ROI for the 50+5 Single Table tournaments?


[/ QUOTE ]

Anything positive (beating the rake). You probably think I'm joking. I am not. If you can get over 10% long term multitabling that would be very nice.

[ QUOTE ]

I have been doing well, and wanted to know if my winrate is sustainable.


[/ QUOTE ]

Given this statment you're winrate is almost surely *not* sustainable. But, no one can tell you your winrate unless you've played a bajillion STTs...what does Eastbay always say? After like 3000 (given that you're playing at the same skill level and some other assumptions) STTs you know your ROI at most within +-5%? Summin like that. And 5% is a lot of ROI.

[ QUOTE ]

How many tournaments do you need before you start approaching something of significance. I assume ROI includes the vig.


[/ QUOTE ]

A bazillion. I'd say 1000 at least but even that doesn't really point very closely at all to any true ROI. You may sorta know statistically if you're winning though.

[ QUOTE ]

thanks for your help.


[/ QUOTE ]

I doubt I did.

[ QUOTE ]

Jason

[/ QUOTE ]

I've always liked this name.....dunno why. I wonce built a big LEGO dragon in like 5th grade and called him Jason....that was one pimpin, bling-blingin, and mack-daddyin dragon lemme tell you. Damn, I wish I kept it or something.....it was certainly the best LEGO thing I eve made. And I could prolly rig him up now so he could breathe fire and sh!t...../images/graemlins/grin.gif

Yugoslav

[eastbay]

Given fixed game conditions:

After 500 games, you have at best a 50% chance of being within +/-5% of your long-term ROI.

After 1000 games, you have at best a 2/3 chance of being within +/-5% of your long-term ROI.

After 3000 games, you have at best a 90% chance of being within +/-5% of your long-term ROI.

eastbay

[Pepsquad]

[ QUOTE ]
Given fixed game conditions:

After 500 games, you have at best a 50% chance of being within +/-5% of your long-term ROI.

After 1000 games, you have at best a 2/3 chance of being within +/-5% of your long-term ROI.

After 3000 games, you have at best a 90% chance of being within +/-5% of your long-term ROI.

eastbay

[/ QUOTE ]

Eastbay,

How did you come up with this? Certainly not qestioning this but here is what I always felt - no mathmatical caculations to back it up - just a gut feeling.

After 500 games, at WORST a 50% chance of being within 5% of your long-term ROI.

After 1,000 games, at WORST a 90% chance of being within 5% of your long-term ROI.

After 3,000 games, not being within 5% of your long-term ROI would be an act of God.

Do you haves links supporting your argument? I'd really be interested in reading more on this.

pep.

[raptor517]

after 5k sngs your roi will be within 10% of where it should be. thats all i know. holla

[eastbay]

[ QUOTE ]

Do you haves links supporting your argument? I'd really be interested in reading more on this.

pep.

[/ QUOTE ]

The method I used to get this is really quite simple. Pick plausible expected values for placing distribution for a 50/30/20 payout game, say 14/13/12%.

The expected value of ROI is then trivial to compute.

Write a computer program to run an ensemble average of N games with this placing distribution, and record how often the actual ROI falls within +/-5% of the expected value at the end of the N game run. Run 500k or so simulations of N games to converge the answer for each N.

Voila.

Vary for a few different ROIs, and notice that the results are quite close for a reasonable range of ROIs and choices for placing distribution. Average some representative results to arrive at the rules of thumb.

The usual arguments against this (skill changes, player pool changes, etc.) would tend to make these estimates too optimistic, rather than the other way around.

eastbay

[Pepsquad]

[ QUOTE ]
[ QUOTE ]

Do you haves links supporting your argument? I'd really be interested in reading more on this.

pep.

[/ QUOTE ]

The method I used to get this is really quite simple. Pick plausible expected values for placing distribution for a 50/30/20 payout game, say 14/13/12%.

The expected value of ROI is then trivial to compute.

Write a computer program to run an ensemble average of N games with this placing distribution, and record how often the actual ROI falls within +/-5% of the expected value at the end of the N game run. Run 500k or so simulations of N games to converge the answer for each N.

Voila.

Vary for a few different ROIs, and notice that the results are quite close for a reasonable range of ROIs and choices for placing distribution. Average some representative results to arrive at the rules of thumb.

The usual arguments against this (skill changes, player pool changes, etc.) would tend to make these estimates too optimistic, rather than the other way around.

eastbay

[/ QUOTE ]

Thank you.