Jason Hahn
@jasonlhahn.bsky.social
No, I don’t think that’s right at all. If it’s shown to be much better than humans then that’s all that’s really necessary for its adoption to be a good thing.
February 7, 2026 at 5:52 AM
No, I don’t think that’s right at all. If it’s shown to be much better than humans then that’s all that’s really necessary for its adoption to be a good thing.
Yeah but how would they know that? It’s not like they got bitten by the same thing 20 minutes ago
January 11, 2026 at 11:27 PM
Yeah but how would they know that? It’s not like they got bitten by the same thing 20 minutes ago
49ers played yesterday so they get minimum a regular week of rest already
January 5, 2026 at 12:57 AM
49ers played yesterday so they get minimum a regular week of rest already
There’s been over 1000 sacks in the NFL this year and like 80 roughing the passer calls.
I think it’s possible to tackle the quarterback.
I think it’s possible to tackle the quarterback.
December 21, 2025 at 8:12 PM
There’s been over 1000 sacks in the NFL this year and like 80 roughing the passer calls.
I think it’s possible to tackle the quarterback.
I think it’s possible to tackle the quarterback.
Odd that next gen stats had conversion probability at 51% (compared to 60% at ESPN) but they marked it as a toss-up on the broadcast.
December 19, 2025 at 2:20 AM
Odd that next gen stats had conversion probability at 51% (compared to 60% at ESPN) but they marked it as a toss-up on the broadcast.
This is true and not at all surprising.
But also it should be expected that it won’t stay true for much longer. Whether that’s 6 months from now or 3 years AI will almost certainly become very good at this in the near future and critiques of this sort should keep that in mind.
But also it should be expected that it won’t stay true for much longer. Whether that’s 6 months from now or 3 years AI will almost certainly become very good at this in the near future and critiques of this sort should keep that in mind.
December 17, 2025 at 4:52 PM
This is true and not at all surprising.
But also it should be expected that it won’t stay true for much longer. Whether that’s 6 months from now or 3 years AI will almost certainly become very good at this in the near future and critiques of this sort should keep that in mind.
But also it should be expected that it won’t stay true for much longer. Whether that’s 6 months from now or 3 years AI will almost certainly become very good at this in the near future and critiques of this sort should keep that in mind.
Check out the link halfway down the article that explains “Total Points”. That’s exactly what they’re doing for all players, offense and defense.
December 3, 2025 at 10:17 PM
Check out the link halfway down the article that explains “Total Points”. That’s exactly what they’re doing for all players, offense and defense.
(I think you just missed that the original post wasn’t serious.)
November 30, 2025 at 3:26 AM
(I think you just missed that the original post wasn’t serious.)
The funny thing here is I think some analytics models might not even agree with going for 2. I know @sethwalder.bsky.social has posted before their WP model sometimes prefers a PAT
November 28, 2025 at 11:22 PM
The funny thing here is I think some analytics models might not even agree with going for 2. I know @sethwalder.bsky.social has posted before their WP model sometimes prefers a PAT
Yes. Dallas over Atlanta in 2020.
Now, how many times has a team that missed the 2-pt conversion on the second TD won? Probably never because it’s too late to do anything about it at that point.
The basic logic is clear although the numbers apparently are a little mixed.
Now, how many times has a team that missed the 2-pt conversion on the second TD won? Probably never because it’s too late to do anything about it at that point.
The basic logic is clear although the numbers apparently are a little mixed.
November 28, 2025 at 11:17 PM
Yes. Dallas over Atlanta in 2020.
Now, how many times has a team that missed the 2-pt conversion on the second TD won? Probably never because it’s too late to do anything about it at that point.
The basic logic is clear although the numbers apparently are a little mixed.
Now, how many times has a team that missed the 2-pt conversion on the second TD won? Probably never because it’s too late to do anything about it at that point.
The basic logic is clear although the numbers apparently are a little mixed.
I think this is fair for a subset of AI (AI generated content) but there’s a lot more to AI than that.
And more relevant is that the generated content will keep getting better to where it isn’t obvious slop. This overarching sentiment you speak of will only apply for a few more years.
And more relevant is that the generated content will keep getting better to where it isn’t obvious slop. This overarching sentiment you speak of will only apply for a few more years.
November 26, 2025 at 5:57 PM
I think this is fair for a subset of AI (AI generated content) but there’s a lot more to AI than that.
And more relevant is that the generated content will keep getting better to where it isn’t obvious slop. This overarching sentiment you speak of will only apply for a few more years.
And more relevant is that the generated content will keep getting better to where it isn’t obvious slop. This overarching sentiment you speak of will only apply for a few more years.
Meh. Maybe I’m not reading it closely enough but other than the silly headline that is walked back immediately I didn’t get the impression from this that the author would be shocked by the Novo Nordisk study results.
November 24, 2025 at 6:57 PM
Meh. Maybe I’m not reading it closely enough but other than the silly headline that is walked back immediately I didn’t get the impression from this that the author would be shocked by the Novo Nordisk study results.
Could a “total” column be added to this for total chance to make the playoffs? That’s the first number I’m looking for at a glance.
November 21, 2025 at 7:59 PM
Could a “total” column be added to this for total chance to make the playoffs? That’s the first number I’m looking for at a glance.
Ok now I’m thinking about a metric that distinguishes between opponent DVOAs (30%, -10%, -10%, -10%) and (-30%, 10%, 10%, 10%).
Feels like the former is more likely to have a better record on average than the latter?
Feels like the former is more likely to have a better record on average than the latter?
November 14, 2025 at 10:31 PM
Ok now I’m thinking about a metric that distinguishes between opponent DVOAs (30%, -10%, -10%, -10%) and (-30%, 10%, 10%, 10%).
Feels like the former is more likely to have a better record on average than the latter?
Feels like the former is more likely to have a better record on average than the latter?
Ten weeks but also because it is (by about 6 days)
November 14, 2025 at 10:24 PM
Ten weeks but also because it is (by about 6 days)
It’s probably a combo of Kyle’s practice and load management, Lynch’s personnel choices, bad luck and other(?).
They should try to make it better, tho, to get the most out of Shanahan’s other abilities. That’s the point of acknowledging it could be something other than luck.
They should try to make it better, tho, to get the most out of Shanahan’s other abilities. That’s the point of acknowledging it could be something other than luck.
November 13, 2025 at 11:25 PM
It’s probably a combo of Kyle’s practice and load management, Lynch’s personnel choices, bad luck and other(?).
They should try to make it better, tho, to get the most out of Shanahan’s other abilities. That’s the point of acknowledging it could be something other than luck.
They should try to make it better, tho, to get the most out of Shanahan’s other abilities. That’s the point of acknowledging it could be something other than luck.
I realize it’s paywalled but this conversation is happening under a link to an article that answers your questions.
Nobody is saying it’s proven. But every year the evidence mounts and the statistically likelihood of random chance gets smaller and smaller.
Nobody is saying it’s proven. But every year the evidence mounts and the statistically likelihood of random chance gets smaller and smaller.
November 13, 2025 at 11:00 PM
I realize it’s paywalled but this conversation is happening under a link to an article that answers your questions.
Nobody is saying it’s proven. But every year the evidence mounts and the statistically likelihood of random chance gets smaller and smaller.
Nobody is saying it’s proven. But every year the evidence mounts and the statistically likelihood of random chance gets smaller and smaller.
The article discusses some, including the level of intensity of Shanahan’s practices. The level of player usage is another choice that could affect injury rates.
Every team has injuries and Bosa or Warner could be unrelated. But the consistently high number means “zero evidence” is wrong.
Every team has injuries and Bosa or Warner could be unrelated. But the consistently high number means “zero evidence” is wrong.
November 13, 2025 at 10:25 PM
The article discusses some, including the level of intensity of Shanahan’s practices. The level of player usage is another choice that could affect injury rates.
Every team has injuries and Bosa or Warner could be unrelated. But the consistently high number means “zero evidence” is wrong.
Every team has injuries and Bosa or Warner could be unrelated. But the consistently high number means “zero evidence” is wrong.
Id disagree with this. The evidence is not definitive, but looking historically the year to year injury rate is not highly correlated so for these Shanahan teams to consistently be the most injured indicates some causation. Could be other things but this is evidence Shanahan’s choices have an effect
November 13, 2025 at 7:45 PM
Id disagree with this. The evidence is not definitive, but looking historically the year to year injury rate is not highly correlated so for these Shanahan teams to consistently be the most injured indicates some causation. Could be other things but this is evidence Shanahan’s choices have an effect
Is DYAR equivalent between rushing and passing such that you could just add them and get overall value?
November 12, 2025 at 8:45 PM
Is DYAR equivalent between rushing and passing such that you could just add them and get overall value?
Something isn’t right there. This is saying win expectancy goes down either way (which shouldn’t happen since the two options you looked at - score or no gain) are the two best possibilities.
November 9, 2025 at 8:00 PM
Something isn’t right there. This is saying win expectancy goes down either way (which shouldn’t happen since the two options you looked at - score or no gain) are the two best possibilities.
Actually I think it might have been from Yglesias where I first saw this brought up. 🤷🏻♂️
October 10, 2025 at 1:40 AM
Actually I think it might have been from Yglesias where I first saw this brought up. 🤷🏻♂️
Not sure about when Biden was in office but lately it’s been clear that Gaza is one area where Democrats could move left in order to court the mean voter so maybe that’s just a topic where “moderating” doesn’t make sense?
October 9, 2025 at 7:16 PM
Not sure about when Biden was in office but lately it’s been clear that Gaza is one area where Democrats could move left in order to court the mean voter so maybe that’s just a topic where “moderating” doesn’t make sense?
Ahh, got it. Thanks
October 3, 2025 at 3:49 AM
Ahh, got it. Thanks
Thanks for posting this, I was curious.
No 4th down page on Thursdays?
No 4th down page on Thursdays?
October 3, 2025 at 3:48 AM
Thanks for posting this, I was curious.
No 4th down page on Thursdays?
No 4th down page on Thursdays?