[Rain Man]

I think cdcox's idea is fine and realistic, and realistic is important since we want to take this to scale.

Another suggestion would be that we know the availability of these players for each week of the 16-week season. Maybe we could use that to define a probability that they'll be available for each playoff week. I'd use a weighted average that increases the weight of later weeks, and then before the season we can roll the dice to define availability, and report those weeks on the roster just like every other week.

If you gave Week 1 a weight of 1 and Week 16 a weight of 16, and ramped up every week in between, I think it would provide a good probability. So if someone misses a couple of weeks early in the season, the odds are good that they'll be in every playoff game. If they missed the last four weeks of the season, they may make it into a playoff game or two, but the odds are that they won't be in all four.

Actually, thinking about the math, I think I'd weight the later games a lot higher than 16. Maybe ramp it up exponentially where Week 16 is worth 30% of the total weight or so. I can come up with a formula to test if you like it.

This system would be nice because it would potentially let some minorly injured players sneak back in for a key playoff game, but probably not all of them.