ITH Poker Forum

[jules_d1]

When playing sng's, what ROI% in considered poor good and great?

[ciaran]

Depends on buy-in, type of SNG, and site.

Anything in the double-digits is quite good in most STTs these days, though.

[jules_d1]

Oh ok. I play Pokerstars $5 and now $10. Mostly 1 table but on the rare occasion i have played 18 and 27 players. Just focusing on Single table at the moment.

$5 SnG
139 games
ROI%- 38%

$10 SnG
47 Games
ROI%- 27% ( Had a horrible value way back in the beginning)

I take it these stats are pretty good?

How many games do you think you need to play and what ROI% before moving up?

Here are my graphs from poker stars. as you can see i used to suck but now my game is much better. Used to play big heads up and such back in the donkey days.

[ciaran]

Moving up is more a function of bankroll, assuming you're winning. I try to stay around 100 buy-ins for the level I'm playing as a minimum, but you can get away with less depeding on your risk tolerance.

ROIs in the $5 and $10 regular tournaments should be higher than in the turbos I play, more so if you're only playing one at a time. A real sample size is more like 1000 games though, so don't rely on your current numbers too much at this stage. And don't read too much into the graphs either. You may have sucked before and gotten better, but it's more likely over a ~200 game sample that you've been about the same player or slightly improving and that you're just seeing normal swings in results.

[Zool1]

I would think 30% is pretty much the absolute cap these days as to what can be acheived by an outstanding player in a regular (non-turbo) structure, one table at a time.

In practice, most SnG players have forgotten what it's like to play one at a time (I'm completely serious), and therefore anything ~20% is considered excellent, but that takes into account some multi-tabling and a turbo structure.

With the number of games you've played, it doesn't tell you much other than it's obviously slightly more likely that you're a winning player than a losing one. As a rule(s) of thumb, I'd say:

- Nothing less than ~500 games at any level counts as a meaningful sample. This figure is not some random largish number thrown about for the sake of it - it's to do with the statistical probability that your results will be within a certain margin of error (caused by short-term 'luck') of where they should be.

- If you play 500 games and have a +ve ROI%, you're probably a winner, although how much of a winner is undeterminable.

- If you play 500 games and have a double digit ROI%, or 1000 games with pretty much any +ve ROI%, you're almost certainly a winner.

- If you play 1000 games with a double figure ROI%, you're almost certainly a solid winner and a really good player.

- You need to play well into the thousands - like more than 3000 - of games to get an idea of what you're actual ROI% is. In practice of course this just doesn't happen - unless you're spacegravy - because people's play changes over this period, they improve, move up, whatever.

GL anyway

And fwiw my own sample is still < 2000 games total across several buy-ins and structures, so other than the fact that I am up overall, I'm pretty much statistically in the dark

[jules_d1]

thanks for the help guys. very insightful.