[msqueri]

Right, but given that “ceiling” isn’t the be all and end all of QB greatness, wouldn’t it follow that you shouldn’t trade a bunch of draft picks to move up a few spots in the draft just to get the “high ceiling” QB?

BV

It seems to me that the argument centers around the fact that determining who is a "high ceiling" QB relies on noisy measurements that have a large standard deviation. One is attempting to maximize the expected value of the total draft. The value of a franchise QB is greater than the value of any individual other player, but possibly not that of several other players. If the draft picks you are giving up are "less noisy" than the one you are getting, then you may be better off keeping them. OTOH, if they are even more noisy (which I think arguably they are) and have a lower total expected value, give them up. Sounds simple, but these expected values are all vectors. What is the correct norm to use in comparing them? Also, some of the effects are non-linear and interacting. If you have a franchise QB and you also have no receivers because you gave up too many draft picks (or paid a veteran QB too much), it doesn't help you nearly as much as his "expected value" would suggest. The evaluation is also a function of time. A rookie may take X years to develop, so you have to integrate over his entire career.

In short, it is a judgement that is neither reliably measurable or well understood. There is also no unique answer (probably) and no way to perform controlled experiments. That is why many more or less reasonable people on this thread do not agree on the answer.

49er draft success since the arrival of Lynch and Shanahan.  Shows that in two measures they were #14 and #12 of 32 NFL teams from 2017 to 2021, but #10 when taking into account positional value.

Riske: A new look at historical draft success for all 32 NFL teams | NFL News, Rankings and Statistics | PFF
https://media.pff.com/2021/02/teams_over...tation.png

This data persuades me that I tend to be too harsh in general with regard to Lynch/Shanahan draft bona fides. It doesn’t tell me much with regard to QB evaluation and drafting, however. They’ve drafted CJ Beathard, Trey Lance, and Brock Purdy. The latter two require many more data points to evaluate. Beathard was arguably a pretty decent pick by 2nd to 4th round QB standards (by production at least, though a skeptic could argue opportunity had a lot to do with that), and for the record I hated that pick at the time so score one for the brain trust. Maybe I should give them some benefit of the doubt in QB evaluation. On the other hand, I think the data I/DC 86/BC have pointed to shows how incredibly difficult QB evaluation is. If Lynch/Shanahan are better than throwing darts that would be impressive. Maybe they are.

[msqueri]

Right, but given that “ceiling” isn’t the be all and end all of QB greatness, wouldn’t it follow that you shouldn’t trade a bunch of draft picks to move up a few spots in the draft just to get the “high ceiling” QB?

BV

It seems to me that the argument centers around the fact that determining who is a "high ceiling" QB relies on noisy measurements that have a large standard deviation. One is attempting to maximize the expected value of the total draft. The value of a franchise QB is greater than the value of any individual other player, but possibly not that of several other players. If the draft picks you are giving up are "less noisy" than the one you are getting, then you may be better off keeping them. OTOH, if they are even more noisy (which I think arguably they are) and have a lower total expected value, give them up. Sounds simple, but these expected values are all vectors. What is the correct norm to use in comparing them? Also, some of the effects are non-linear and interacting. If you have a franchise QB and you also have no receivers because you gave up too many draft picks (or paid a veteran QB too much), it doesn't help you nearly as much as his "expected value" would suggest. The evaluation is also a function of time. A rookie may take X years to develop, so you have to integrate over his entire career.

In short, it is a judgement that is neither reliably measurable or well understood. There is also no unique answer (probably) and no way to perform controlled experiments. That is why many more or less reasonable people on this thread do not agree on the answer.

49er draft success since the arrival of Lynch and Shanahan.  Shows that in two measures they were #14 and #12 of 32 NFL teams from 2017 to 2021, but #10 when taking into account positional value.

Riske: A new look at historical draft success for all 32 NFL teams | NFL News, Rankings and Statistics | PFF
https://media.pff.com/2021/02/teams_over...tation.png

This data persuades me that I tend to be too harsh in general with regard to Lynch/Shanahan draft bona fides. It doesn’t tell me much with regard to QB evaluation and drafting, however. They’ve drafted CJ Beathard, Trey Lance, and Brock Purdy. The latter two require many more data points to evaluate. Beathard was arguably a pretty decent pick by 2nd to 4th round QB standards (by production at least, though a skeptic could argue opportunity had a lot to do with that), and for the record I hated that pick at the time so score one for the brain trust. Maybe I should give them some benefit of the doubt in QB evaluation. On the other hand, I think the data I/DC 86/BC have pointed to shows how incredibly difficult QB evaluation is. If Lynch/Shanahan are better than throwing darts that would be impressive. Maybe they are.