...and I think I'm going to need some help with the math here. I've been trying to see if I can get a formula, but am getting increasingly confuzzled.

Let's use some simple strength ranges of [0,1.0], [0,.80], [0,.50], and [0,.25] - i.e. mean strengths of 50%, 40%, 25%, and 12.5%. (let's call them names Link, Strong, Midcard, and Jobber). The basic assumption for heads-up against Jobber is that Link's realized strength (for a particular voter) will be between 0 and .25 25% of the time, and above it 75% of the time; and that if his strength is below 25%, he has a 50/50 shot of getting that vote. Thus, his expected vote %age against Jobber is: .25*50%+.75*1. Which is 87.5%.

But how does this work 4-way? The thing that came to my mind was to split the first 25% of votes 4 ways, then the next 25% 3 ways (since Midcard is 25% stronger than Jobber), then the next 30% 2 ways (Strong is 30% stronger than Midcard), and give the last 20% to Link. This gives an estimate of .25/4+.25/3+.3/2+.2~=49.6% for the other 3 characters, and thus 50.4% for Link. But Link's simulation based percentage is about 54.6% (and 34.6% for Strong, 10.1% for Midcard, and 1% for Jobber). The estimate is reasonably close, but clearly wrong.

Anyone have any thoughts? The concepts and algebra are a bit tricky here.

Maybe trying this in a simpler 3-way example will generate some insights...?
---
www.gamefaqscontests.com
www.gamefaqscontests.com/gallery