So I am running at 10.29% ROI after 683 games on 3$ and dabbed here and there with the 5$ ones with a 70-90 BI management, but never managed to get into the green.

Now I just had 500$ on the roll which is enough for the 5s. Started a session 3$s and 5$ and run down to 377$ with me having -5.87% ROI after 253 x 5$ games. The sample size of the 5s shouldn't say that much but I still feel the missing money. Now I just have 70 BI left, so I am gonna skip them again.

I was thinking about playing 3$ Turbo as my BRM of 85-125 BI says it's totally ok. On the other hand my ROI at these ones -0.5% after 737 games, which makes it feel so much like a waste of time and my concentration for other games.

Looks like I should only play 3$ reg but getting in big volume with only that stake is hard I guess.

Don't know where to go from here.. I am pretty sure I am reading to much into the samplesize, but I am virtually daily on pokerdope calculating all the **** just to ease my mind, but it works less and less somehow. Also I don't think that everybody on there is an ICM beast or that I lack focus (no tv, no music), but I can't put my finger on it.

thx 4 advice

Did you reduce your number of tables when you moved up in stakes?

I played all games mixed (3$ reg, 3$ turbo, 5$ reg) with a 12-15 table cap with 0-3 mtts added.

Maybe reducing tables to 9 would help, but I am kind of wanting to get enough volume in to be at least at a whimsical hourly (which doesnt work if your roi is negative). It worked getting results in with 3$ and not with 3$ Turbo, but I can't say if it's the table count or only variance.

Don't discount how different the games play from turbo to non turbo.
Sure poker is still poker etc, but your opponents and more importantly the reasons why they play poker will be quite different.
That said, if your ev over 700+ $3 turbo sng's is B/E then there is certainly a fair amount of strategy you can improve on/spots being missed etc.
Your opponents must be pretty horrible bad, so you probably aren't playing in a way to take full advantage.
I can take a look at a HH or two if you want some feedback

This sample isn't nearly large enough to determine true winrate. An EV ROI over that number of games would be better, but still not enough.

In total I counted like ~2000 games and I've got samples of losing/breakeven much longer than that with a ~4% EV ROI (Hypers) and 5-8% (Turbos).

If those ROIs truly reflect your winrate (and even if they don't, I suppose) the biggest general leaks I'd recommend working on/looking at would be:

- Correct push/fold (Estimating calling ranges of our opponents, not the ranges we think they should call with)
- Improving HU play (Drastically affects winrate considering you're playing for 20% of the prize pool)
- Postflop (Getting max value, using correct sizings)

I agree the sample size is way to small to draw any conclusions from your results OP. Recommend that you post some actual hands that you have questions about that we can give you feedback on.

The problem with non-winner-take-all tourneys is that playing the Nash strategy isn't even always profitable, let alone optimal.

One of the most obvious examples is that if, on the bubble, CO and BTN with tiny stacks fold, you're in the SB with a big stack and the BB also has a big stack and is known to call much wider than Nash, you shouldn't shove any two - by calling too wide, BB will be hurting his own equity a lot but will also be hurting yours, to the benefit of CO and BTN.

The definition of the Nash equilibrium is that, if all the opponents use their Nash strategies, then a player can't gain equity by deviating from his own Nash strategy. It doesn't apply at all to the case when some of the opponents deviate from Nash.

So, in games with bad opponents, you have to seek exploitative strategies. Sometimes exploitation consists in folding.

Don't take $0.50 36-mans that seriously I know, they're about the only way to make use of those promotional bunches of free $0.50 tickets, but wasting precious focus on them costs us more than playing them in a push-or-fold fashion from the very start Also, you can play them in the background when you're doing off-the-table stuff like posting on 2+2.

I wasn't aware that using ICM or Nash calculations doesn't allow you to set people with certain ranges...

I'm trying to pretty much do the same thing. I deposited $500 on America's Card Room in June. I've played some MTT(mostly break even), some cash(down $300) and 6 and 9 player turbo SNG.(profitable + rakeback + weekly bonus money).

I have been using a HUD since February. I have built my $500 deposit up to $900 in 3 months. $5 and $10 SNG have been nearly all of my profit. They are the softest online poker games, IMHO. I'm getting ready to move up to $15 SNG, and I would have approximately 60 Buy Ins. I am beating the cheaper games, but I expect the $15 dollar SNG to contain tougher players. Hopefully, I can continue to beat the game while playing at the $15 SNG level. If I have trouble, I would probably look to move back down if my Bankroll drops back down to $650 or so.

I think that I'm taking a rational line. I typically play up to 6 tables at a time. My HUD stats are 26 VPIP, 18.6 PFR, 6.9 3-Bet in 178,360 hands at WSOP NJ and 22.1 VPIP, 16.2 PFR, 6.7 3-Bet in 47,376 hands on America's Card Room. My HUD helps me to identify Villains who are exploitable. On WSOP NJ, I found that only 2 Villains played extremely well, to the extent that I had difficulty exploiting them. On ACR, I have ran into some tough regs, but no one has really given me problems at the $5 & $10 level.

Any thoughts?

You should think through how many concurrent tables you play, there is no set amount that suits all but some people get confused over the variance part, often they think if they play lots more tables they smooth out their results - sadly it is often the opposite.

Here is a little example to explain some of what happens.

The variance per game is roughly the same for any practical roi but the amount of games per time period is a big factor of variance.
eg, for 9 seat STTs with approx 10% rake variance per game is about 2.2 (in BBI^2).

The variance for 789 games is 789 x the variance for one, ie, n x var per game.

So lets say in 1 year player A plays 10000 games at an roi of 2% and player B plays 5000 games at an roi of 4%. Both players on average pick up the same in a year.
A gets: 10000 * 0.02 = +200 BIs
B gets: 5000 * 0.04 = +200 BIs

Which one has the smoother usually easier time? It's player B

Variance seen for A in year = 10000 * 2.2 = 22000, so the standard dev = sqrt(22000) = 148.3

Variance seen for A in year = 5000 * 2.2 = 11000, so the standard dev = sqrt(22000) = 104.9

So ~68% of the time the amount A gets is obtained from a bell curve centered on 200 BIs but +/- 148.3BI's

So ~68% of the time the amount B gets is obtained from a bell curve centered on 200 BIs but +/- 104.9BI's

This is a pretty bad amount of uncertainty added but even worse than this is what the typical largest downswing seen is.
Typically 50% of years player A, playing 10000 games, will see a 112 BI downswing somewhere whereas B, playing 5000, 50% of years only sees a 70BI downswing. The length of time digging out of these swings is also really bad for A, they play twice as many but earn only half so take longer to get out.

Overall I would recommend playing the amount of tables that makes you feel comfortable, earns a good deal and also importantly allows you to learn very fast so you can maybe jump up levels.