ROI question

[thew92]

I have started playing Sngs 2,3,5 and 10 tables at poker stars. So far I have a total of 90 SnGs under my belt. I have a ROI of 51%. How many SnGs do I have to play before I can start to trust my ROI. I believe in limit hold'em BB/100 doen't mean any thing untill you have at least 10k hands.

[jmbreslin]

Who knows? Some people say you have to log several hundred at the same buy-in to overcome the impact of variance, but I've found that I can get a pretty reliable read on my ability at a particular level after 100+. I think I played about 170 at $1.20 before moving up when I first started, but my performance was pretty consistent well before I reached that point.

What SnGs are you playing?

[Simpioni]

Strict maths experts wouldnt trust a sample size under thousands and thousands and thousands!

Its better not to focus on that and focus on your decisions IMO. If you consistently get your money in good then your doing well. If youve caught yourself getting your money in bad on more than a few occasions but outdrawing, then you need to learn from these experiences as if u had lost.

[Schlepper333]

Some say you can never get a true read on ROI, but most articles I have read on the subject indicate that 1000 games is standard.

Check out:

http://www.internettexasholdem.com/phpbb2/acceptable-roi-levels-by-sng-level-vt47544.html

[thew92]

[ciaran]

The confidence calculator here, http://www.aleomagus.freeservers.com/Spreadsheet/, will give you some good numbers for single table games. It could probably be modified for larger games (certainly the 2/3/5 table games) but I'm not up for it.

I think the standard deviation should grow as the size of the tournament increases, but the maximum potential ROI should as well. Still, I think you'd need at least several hundred at a minimum to get a decent idea of what your true ROI might be. Note that in the included example in that spreadsheet a 23% winner in single-table games is still only 99.5% sure to be a winner over 300 games, and 23% is a very high ROI for single-table games these days.

At the lower levels, I think it would be difficult to have a clue (that is, if you actually study the game at all) and be a loser, so the advice to make sure you're always making the right decisions and not focus on the actual ROI is probably good. At higher levels, it will become more important to figure out if you're good enough to beat the games at all.

[dktoller]

Couldn't find your link ciaran, but the standard deviation definitely grows with the tournament size. Assuming you win at a rate of +30% (pre-rake) and your itm placements are evenly distributed, then:
[Edit: 30% "pre-rake" corresponds to 18% on a post-rake basis, which is the way ROIs are normally reported.]

9-man: mean = 1.30, sample std dev = 1.80
18-man: m = 1.30, s = 2.31
45-man: m = 1.30, s = 2.68
180-man: m = 1.30, s = 5.22

Going at this clip for 130 such game for the 9-man tables, you could state with ~90% confidence that you are a long term winner (over the 10% rake).
(Note this is still a pretty vague estimate of your ROI, it just says that it's likely greater than zero.)

For 180-man tourneys you would need 1200 games to make this same statement, due to the much large variance. Of course, if you do better than 30% fewer games would be needed.

This results-orientated approach doesn't converge quickly enough to be of much use for the majority of players, IMHO.
Posting hands, sharing hand histories, and generally analyzing your game for leaks are much more productive in determining whether you are playing well.

[ciaran]

Fixed the link in my original post.

And if you can win at a 30% clip in 9-man SNGs, you need to teach me how.

[Evilrulz]

Ive averaged between 30-35% ROI on 9 man 10$ SNG's doing 9 tabling. I have about 300 done and going to do somewhere between 1-1.5k but I dont see my ROI doing much changing. As I get more and more done it just takes too huge a swing for it to move in either direction. There reg speed but the ROI more than makes up for it $$$/hr wise.

[ImBetterDude]

215% at the 5 table ROI is skewing your results. It seems you've had 2 or 3 big cashes, which is good...but that kind of results you shouldn't expect as regularly as they have happened thus far.

12% on 2 table turneys should be more telling.