Hi all

I was hoping someone could provide some info on sample sizes, specifically when is a sample size big enough to establish a true roi within defined deviations?

For example, a 2k sample size and a 40% roi, would there be a deviation of x% either way which would likely play out over a larger sample size (assuming initial data has not been skewed by some large score in the first instance)?

Any help appreciated

Are we talking about big field MTTs or 180s? I'm not sure what % of deviation you'd get but I think finding your true ROI in MTTs will take 10s of thousands of games, but if your making correct plays then don't worry about your roi as much and keep making that loot.

Thanks Russ.

Was refering the scheduled mtt's as opposed to 180's/sng's. I remember seeing a HU chart once which gave deviations based upon volume played, thought there might be an equivilent.

Just how many thousands of games is enough to get a 'true' approximation?

Would 5k games yield any significant data?

As you pointed out just making da monies is what is really important, yet I was thinking about potentialy exploring staking deals for bigger games in the future and wondered what kind of stats I would need to generate interest.

Thanks again for your help.

It's really hard, because your large field MTTs consist of different buyins and field sizes; obviously, your ROI will differ among them.

So, even if it were possible to calculate ROI, it would be easier to do so for specific MTTs I think.

or if you are any good with spreadsheets I think splitting the games up by BI (e.g. 1$-5$ 5$-10$ etc) and runners to get a larger sample per BI. Would be good for what you are aiming for imo

*edit*
I dont think you can ever get a sample large enough tho if you are splitting up max 5k games into BI's

Thanks leviathan

The schedule I play is very consistent, with a lot of the games around the same buy ins, though field sizes differ.

This is the rub of it, I don't need or have to consider staking for the games I currently beat up, yet even with a good roi (40% would be consider good amiright?) these games will only yield so much. I am ready to play higher.

How can you put yourself out there to establish sound results in the games you want to play if you don't have any sample size of any significance in those games?

Big shout out to the 2 + 2 mtt community for helping a noob to the forum.

Many thanks Yakmelk

[ ] Good with spreadsheets .

I rely mostly on opr/sharkscope for my data.

Your prolly right about splitting the buys in's but as per your edit, 5k games dosent seem that much to work with then.

So probs best to keep on trucking till I get a bigger sample size then? Which kind of brings me back to my original point of how many games are enough?

Perhaps considering it another way might help. So if you were in the market to stake someone, what kind of data would you want to see to get on it?

Problem is, lets say 15,000 is the desired number to accurately predict your "true" ROI. By the time you reach 15,000 the games(style of play) will be a lot different then they were 15,000 games ago. So now you'd have to figure out your new "true" ROI for the new type of playing style.

Just play your best...and what?...forget the rest.

Interesting stuff twonine, thank you.

So you can never truly know and your roi is pretty much irrelevant? So how can you know that your just not running good/bad over a protracted period of time?

From a staking perspective, is it mostly reputation that leads to success, or more a case of knowing someone personally who would back you?

So building a reputation is something I should work on? I've always tried to stay anon previously.

Thanks again.

my roi has gone down 20+% over the past year and i have an ABI of 50-60. I think I'm sucking more and running worse! BTW, I play like 1500 or so mtts/year on each of two main sites :S

1500 games is hardly enough of a sample to worry about a 20% spread. u could be a farrrrrrrrrrr better of a player compared to the fields and still be making farrr less $$ and have a lower roi over such a tiny sample.

People adapt to your game over time too...

you presumably fix random leaks in your game here and there so your ROI is hopefully improving over time as well...

if you have a few thousand games i'd be willing to say you're probably within +/-25% of your true ROI.

it's really truly impossible to know tho... if you get your money in good 90% of the time and make really well timed moves your ROI will be high, otherwise it will probably stay between 10-40% where you're at now i'm guessin'.

a reputation and lots of posts on the forums etc def help for getting backing. winning some name brand tournaments helps too.

you could always try a backing arrangement that favors your backer at the beginning then shifts towards 50/50 as you improve / as your sample size increases?

OP, if you're looking to get staked go check out the marketplace and start a thread about why you want to get staked, give links to OPR and scope and let people decide for themselves if you're worth staking.