SNG results at ACR and BOL

[NWPatriot]

I play at both ACR and BOL. I have played a few hundred SNG's and I am seeing an alarming disparity between my results at these two sites. I have a +18% ROI in SNG's at ACR but a -11% ROI at BOL.

I get that variance is a real thing, but I wondered if anyone else has noticed this. I have also confirmed in my own hand history data that I win way more with lesser hands at BOL than I do at ACR (like small pairs for instance, fair a lot better for me at BOL - this means they may do the same for others as well).

To me there are 3 possible explanations:

1. The hand distribution at BOL is skewed towards lesser hands winning somewhat more frequently.
2. Players at BOL are more likely to play worse hands so when they hit they are more visible.
3. Players at BOL are better than I am and I can't figure out what the heck I am doing wrong.

Of course, there are other options. Curious about your thoughts.

Good luck and God bless.

 

[RicardoInciarte]

Hi that is an interested topic once when I played at BOL (just comming basck to poker and Cardschat happy about it ) I felt kind of the same thing but now I won't play on BOL I will put my focus more on Stars and ACR are my favorite sites I'm playing now more on tourney nad not much SNGs... Not sure about you theories but I feel like the 1st is more true xD ... good luck !

 

[NWPatriot]

Thanks

Not sure how many folks have data at both. Just trying to support my data with others' data.

 

[fundiver199]

Assuming, that you have played around 100 SnGs on each site, then this is within completely normal variance. Assuming standard 9 man SnGs the 70% confidence interval for a break even player after a sample of 100 is -15% to +15% ROI. This mean, that 15% of the time, you will do worse than -15% ROI, and 15% of the time you will do better than +15% ROI. Even after a sample of 1.000 the 70% confidence interval is still -5% to +5%. So basically "the long run" is just a lot longer, than most of us tend us instinctively assume

https://www.primedope.com/tournament-variance-calculator/

 

[NWPatriot]

Assuming, that you have played around 100 SnGs on each site, then this is within completely normal variance. Assuming standard 9 man SnGs the 70% confidence interval for a break even player after a sample of 100 is -15% to +15% ROI. This mean, that 15% of the time, you will do worse than -15% ROI, and 15% of the time you will do better than +15% ROI. Even after a sample of 1.000 the 70% confidence interval is still -5% to +5%. So basically "the long run" is just a lot longer, than most of us tend us instinctively assume

Thanks for this perspective. Yes, I had assumed that 200 SNG's (100 each) was sufficient to do some comparison, I had dismissed variance as an explanation, because these two datasets seemed too far apart.

I see that you must have input 0% ROI (breakeven player) into the variance calculator in order to generate the numbers you stated. I hadn't thought of doing that, as every time i had used the calculator, it was more to validate my own personal ROI and see how good or bad it could/should be. This is a really good way to set a baseline. I would have input my actual 18% ROI to see how good or bad it could be for the next 100 (answer is 3% - 33%). Knowing this, the BOL info is outside of what my actual ROI might suggest.

So I guess it is all about how we use the tool. Your 0% ROI input does make good sense. I will need to consider how best to use the variance calculator.