[BvlyHls90210]

I am an 8 tabling 109 speed SnG regular. Let's say over my last 1000 SnGs I have an ROI of 7.5%. Now I also play the 109 regulars and the 77 speeds from time to time for various reasons, but my stats in those games are much less valid due to sample size. In any case, during my last 1000 speed 109s, I have played 100 77 speeds at an ROI of 15%, and 100 regular 109s at an ROI of 22.5% .

My goal is to maximize my hourly rate. That means if I am to switch to playing 77 speeds, my ROI would need to be 41.5% higher, and if I were to switch to the 109 regulars, my ROI would have to be 100% higher (assuming they take twice as long). Based on my current sample, these conditions are being met, and if my sample was valid, I am leaving money on the table by playing my least profitable game).

So my question is at what ratio of my 109 speed sample size to the other samples can I validly make these conclusions? Or ]am I already there? I know 100 means absolutely nothing, but I feel like someone who knows a lot about statistics and poker could say the 100 regular 109 SnG results do mean something if we know with some statistical validity what the same person is doing at speed 109 SnGs.

Tangentially I'd be curious if anyone has big sample like these that can help me with these comparisons (between levels and between regulars and speeds at the same level).