Okay, I'm retarded, I just did some quick math, and it totally didn't work at all. I was trying to quantize the stats to a maximum of one. But dividing by games played is the completely wrong way to do it. That's already taken into account by the fact that you're dividing. A team that's played 5 games could have the same ratio as a team that's played 8 games. That ratio just needs divided by the current maximum ratio. I guess the alternative is to subtract the bad factor from the good, then divide by games played, then quantize. Maybe I'll try both and see which I prefer.
Update: I knew there was a reason I did a ratio instead of an average difference. Losing teams will likely have negative ratios for their stat components. That would mean you could have a scale from -1 to 1, which woudl mean adjusting everyhting to be positive, then re-quantizing. No thank you. I'm going with my first intent, the quantized ratio method.
Update: I knew there was a reason I did a ratio instead of an average difference. Losing teams will likely have negative ratios for their stat components. That would mean you could have a scale from -1 to 1, which woudl mean adjusting everyhting to be positive, then re-quantizing. No thank you. I'm going with my first intent, the quantized ratio method.
