09-05-2017, 08:28 PM
I'm saying that I'm maybe willing to move a bit more from my priors than you are. The "begging the question" aspect of this is that as you are describing it,m it appears your priors are coloring your interpretation of the observed result (a form of confirmation bias), although maybe you are describing more how the data is affecting your posteriors.
In football it is a given that there are occasionally fluky plays that affect the outcome or closeness of a game, and players are temporarily injured or unavailable or what have you, which is why it is reasonable not to move from one's priors based on a single result (especially one that didn't dramatically differ from the expected; it's not like WMU upset USC). But I think if you try and pick and chose which plays are fluky and which ones support your argument, you can get into trouble. As others have pointed out, why do you discount the KOR and a trick play by WMU and not the pick six and missed tackle leading to USC touchdowns. Maybe the fact that USC gave up a kickoff return for a touchdown means that their special team play is substandard and that we might have several big returns Saturday. Maybe on the trick play, it's a sign that USC overpursues and that they are vulnerable to giving up a big play in a similar situation to us (we have been known to punish teams with trick plays when they sell out for the run).
(https://www.youtube.com/watch?v=H3V93JHehrs)
And of course if the dropoff in defense is so pronounced that one player was key, it may means there is a big fall off from that player to the back-up and given the non-zero probability of injury or of said player being ejected (again), that too bodes well for us. Or maybe it is a bit of a fluke, and USC will revert to expected against us (I hope not). I don't know that all of that is enough for me to conclude that we are going to crush USC like we crushed Rice, but subjectively I feel better this week than I did before week 0, both because we had a very good game and because USC had an ok game, and I think the data does support moving the posteriors a bit from the priors.
BC
In football it is a given that there are occasionally fluky plays that affect the outcome or closeness of a game, and players are temporarily injured or unavailable or what have you, which is why it is reasonable not to move from one's priors based on a single result (especially one that didn't dramatically differ from the expected; it's not like WMU upset USC). But I think if you try and pick and chose which plays are fluky and which ones support your argument, you can get into trouble. As others have pointed out, why do you discount the KOR and a trick play by WMU and not the pick six and missed tackle leading to USC touchdowns. Maybe the fact that USC gave up a kickoff return for a touchdown means that their special team play is substandard and that we might have several big returns Saturday. Maybe on the trick play, it's a sign that USC overpursues and that they are vulnerable to giving up a big play in a similar situation to us (we have been known to punish teams with trick plays when they sell out for the run).
(https://www.youtube.com/watch?v=H3V93JHehrs)
And of course if the dropoff in defense is so pronounced that one player was key, it may means there is a big fall off from that player to the back-up and given the non-zero probability of injury or of said player being ejected (again), that too bodes well for us. Or maybe it is a bit of a fluke, and USC will revert to expected against us (I hope not). I don't know that all of that is enough for me to conclude that we are going to crush USC like we crushed Rice, but subjectively I feel better this week than I did before week 0, both because we had a very good game and because USC had an ok game, and I think the data does support moving the posteriors a bit from the priors.
BC