**I am a bit unsure what to mark this as/title this as. We refer to this type of phenomena as 'volatility' but this apparently has a specific context with regards to statistical phenomena so any suggestions to that will help me as well.

I was wondering if anyone knew the best way to quantify the degree of variability for the following scenario. Suppose we are taking a snap shot of the top 3 best teams in a league of 30 teams every week for 20 weeks.

For example

Week 1:

1. Team A    0.673
2. Team B    0.648
3. Team C    0.637

Week 2:

1. Team D    0.713
2. Team A    0.642
3. Team E    0.596

Week 3:

1. Team B    0.643
2. Team D    0.618
3. Team F    0.598

...

Week 10:

1. Team F    0.613
2. Team C    0.602
3. Team G    0.598

We only know the top 3 teams and cannot infer anything about the rest of the standings. Additionally it would be interesting to see some sort of relation to Top 3 Win % but this isn't a necessity. What would be some methods to calculate the variability of the top 3 positions? In order to potentially offer forecasts for future weeks.