My take:
Highly probable wins (prob. of victory > 0.900 for each game): EIU, BSU, IU, ISU, MINN
Likely wins ( 0.700 < prob. of victory < 0.900 ): NW, UA, MSU, MICH
Borderline wins (0.550 < prob. of victory < 0.700): UW, PSU
Push (0.450 < prob. of victory < 0.550): tOSU
"BY THE NUMBERS" worst case: (Calculating using the low range probability values as our null hypothesis)
expected # of victories = 8.85 games
standard deviation = 1.43 games
"BY THE NUMBERS" best case: (using the high-range values for our null hypothesis)
expected # of victories = 10.55 games
standard deviation = 1.01 games
Now using a 2-standard deviation "rule of thumb" in order to discern what should constitute a "usual" outcome for each null hypothesis I arrive at the following results:
Worst-case: Here the "usual-range" would be between 6 and 11.7 wins with a mean outcome right in the middle. This obviously reveals BOTH the lack of resolution that you can have in this sort of analysis AND the REALITY that a team like Iowa really CAN have a VERY narrow margin for error.
Best-case: Here the "usual-range" would be between 8.5 and 12 wins, with a mean outcome at right around 10.5 wins.
Conclusion: The reality in 2010 probably lies somewhere in between the worst and best cases. I'm inclined to bet that a mean of 9.75 wins with a standard deviation of 1.25 games is probably the closest thing to reality for the 2010 Hawk squad. That would imply that an 11 win season would simply be a single standard deviation away from the mean ... making it a VERY attainable outcome. However, nearly equally likely would also be an 8.5 win season (roughly speaking 8 wins with a game right on the border that could go either way).