[BostonCard]

As someone who believes Costello to be superior, my worst case scenario is both players getting a few possessions and Chryst plays at the peak of his ability and Costello has an off day, cementing Chryst's place as QB1.

Sounds a bit like epistemic closure.  An alternative way of looking at the same data is that Chryst is actually a better QB than Costello, and your assumption that Costello wsa superior is wrong.

I am inclined to agree with you that Costello is superior.  But if Chryst and Costello were to both get a few positions and Chryst were to play better, I would re-examine my assumptions.  It would probably be too few datapoints to conclude, categorically, that Chryst was the better QB, but it would certainly lead me to question my a priori assumptions.

The best QB given their respective performances vs Utah = The probability of observing their performances against Utah given their respective quality x Their underlying quality

(Or, posteriors = priors x data)

BC

[BostonCard]

As someone who believes Costello to be superior, my worst case scenario is both players getting a few possessions and Chryst plays at the peak of his ability and Costello has an off day, cementing Chryst's place as QB1.

Sounds a bit like epistemic closure.  An alternative way of looking at the same data is that Chryst is actually a better QB than Costello, and your assumption that Costello wsa superior is wrong.

I am inclined to agree with you that Costello is superior.  But if Chryst and Costello were to both get a few positions and Chryst were to play better, I would re-examine my assumptions.  It would probably be too few datapoints to conclude, categorically, that Chryst was the better QB, but it would certainly lead me to question my a priori assumptions.

The best QB given their respective performances vs Utah = The probability of observing their performances against Utah given their respective quality x Their underlying quality

(Or, posteriors = priors x data)

BC