Why the 49ers should not worry about tomorrow, or the future with Nick Mullens if Garoppolo misses time. Diving into the statistics + a shocking revelation about CJ Beathard - why Shanahan kept him.
This is a really interesting article. And the statistical analysis, to me speaks to a key point of consideration: the distinctions between qualitative and quantitative analysis. I think it’s overly simplistic to say CJB’s receivers were bad and Jimmy’s were good. Or that BDN was luckier with the same cast (or something to that effect). I think part of the distinction is the QB’s “understanding” of his receivers. Not just getting the balls where they “should be”, and where they are “catchable”. We often hear Jimmy speaking of how he understands how his receivers are running their routes etc so his passing is “nuanced” (and this would be the qualitative difference that I use to “explain” the gulf in the difference in the rates of completion—Jimmy doesn’t try to pass “to” his receivers, but “for” them—this might also explain Jimmy’s reluctance to pass to Pettis). So while CJB can “make” the throws, he is not “passing” the ball with the same quality as Jimmy.
Anyway, that’s my speculation. It’s not as simple as Jimmy had better receivers. Jimmy contributes to making his receivers better. CJB hasn’t done that ... yet.
absolutely. Jimmy makes his WRs better which creates the illusion that they were already good. but its interesting to see that CJB's weaknesses might not be weaknesses. Garoppolo is a more advanced QB but I think beathard has something left.
Really good article to read Riqo, thanks. My aside about Pettis, maybe he runs his routes well, but is just a little “skatty” (or shaky) in how he’s finishing them off so Jimmy just doesn’t quite “understand” where he needs to put the ball for him. That pass in the Cards game, Shanny noted Pettis got bumped off his line early in the route. Another breakdown vid (johnnydels) said Pettis ran a slighty shallow angle in the deep route that left him short of where he “should” have been. So it’s like Jimmy’s passes to Pettis are basically hail mary’s—he puts them where they’re supposed to be and hopes Pettis gets there...
There are lies, damn lies and statistics. CJB just doesn’t pass the eye test in a game.
I get why we have 3 QBs, if it hadn’t been a Covid year, I wonder if Kyle would have kept Broc to bring along and lose CJB as he couldn’t beat out Mullens.
That's what im saying though - i was a staunch supporter of cutting beathard prior to this season. IMO hes a bust, and a big one - but this made me rethink things. Shanahan obviously knows more about this offense than we do. And although Mullens has outplayed Beathard, Shanny referred to the two as equal during training camp. so maybe he sees what these stats also see in Beathard - that he isnt as bad as we think. Food for thought. I still think Mullens is getting traded once Jimmy is healthy - he hits FA after this season and if the 49ers can get a pick from a desperate team, theyll pull the trigger.
No one is going to trade for Mullens. There are a lot of Teams who may have similar backup QBs. Just remember Shanny's scheme and playcalling makes it a tad easier for the QBs to look competent, if you have a minimum skill-set
True but teams get desperate and injuries change things. In 6 weeks is the trade deadline - lets say Steelers are contending and Big Ben goes down - they won't play Rudolph or Hodges who were abysmal. They'll be interested.
I still don't get keeping 3 QBs, even this year. If your starter goes down, you can always get someone off the street to backup the 2nd stringer. If he goes down, your season is toast anyway.
I think last year we kept 3 as Jimmy was coming off an ACL. Made sense.
This year we're coming off a superbowl run, we're in the "window", so we've kept 3 to maximise chances of staying competitive if Jimmy went down for a period of time.
I might be misunderstanding your analysis, but if I read it right you've fundamentally misunderstood the expected completion stat. Apologies if I've read your analysis wrong.
Expected completion is what you'd expect an average QB to complete based on the situations he's been in (taking into account distance, seperation, pressure etc.). It doesn't mean "what you'd expect the WR to catch given how well he threw the ball".
In other words, if I was the QB and attempted the same throws as CJB, then my expected QB would also be 68% - because it's not measuring how good my throws were, it's measuring how hard the situation was.
So Beathard's 68% expected completion stat does not mean he did his job well - it means he was put in situations where you'd expect an average QB to complete passes 68% of the time.
In Shanahan's offense, that number is always high - he puts his QBs in situations to succeed. So when you say "best in the league", that's misleading because it suggests CJB was achieving it. A better description would be "easiest throws in the league" - and yet he still wasn't succeeding.
The bit you are attributing to the receivers - CJB's 8% below average - is the CPOE stat, and is a pretty useful QB stat (and not a useful receiver stat).
This matches the eyeball test. Shanahan's scheme wasn't any worse with CJB on the field - the QB was.
Might be a different stat you are looking at. my stats are derived from NGS - they are defined as higher quality = higher xCOMP vs. lower quality throws = lower xCOMP.
Try that ^. but from what I understand - a higher xCOMP is derived from the NGS completion probability stat and so that would mean whatever the xCOMP% is = the amount of high quality throws they made, regardless of completed. So the differential - that would be how the actual compared to what should have been completed (by the Qb and WR) - in this case, CJB being so high in that category means that he made a high percentage of quality throws (based on their NGS completion probability) - and the differential means that he got unlucky on those actually being completed
Yeah, you have to read carefully what the stat actually does. I'm very much in the "pro-stats" camp, but you have to be careful to understand what the stat is actually measuring (because the main reason we need next-gen stats in the first place is that last-gen stats are totally confused as to whether they are measuring player performance, coaching performance, team performance etc.).
[Or you can not to that. It's worked well for David Lombardi....]
"Expected Completion Percentage (xComp) gives an indication of the level of difficulty of a quarterback's throws. A high xComp indicates a quarterback's propensity to throw higher probability passes, while a low xComp tells us a quarterback has a tendency to take more chances."
Right, but "high probability pass" means a safe pass, not a good pass.
See the preceeding sentence to the one you quoted at the link below "Expected Completion Percentage (xComp) gives an indication of the level of difficulty of a quarterback's throws."
It's a measure of difficulty, not a measure of quality.
Yeah looks like you might be right. This is my first stuff with advanced stats so still learning 🤣 - but this also means a couple of articles from big outlets ive read are wrong too. Weird, but yeah you are right. Their explanation is super weird, at least to me. Edited (but the aggression and LCAD still stand.)
"Completion Percentage Above Expectation (xComp +/-) measures a quarterback's performance relative to the difficulty of their throws on an aggregate level."
The stat that CJB sucks at is the one that measure his quality. The one he is "best" at is the one that measures how easy his passes are (i.e. essentially the one that measures Shanahan's quality!)
Yeah, that sounds logical regarding Completion Percentage Above Expectation. If you are completing passes higher than the expected figure then you're probably good to very good.
I could be wrong. And even if I am - your point stands - Shanahan consistently makes his QBs make easy throws, which is why Mullens, Jimmy G, and Beathard all are similar in this department. But its interesting.
i think you might be right. NGS did a terrible job of explaining it. but the probability of completion stat makes me hesitate - "The probability of a pass completion, based on numerous factors such as receiver separation from the nearest defender, where the receiver is on the field, the separation the passer had at time of throw from the nearest pass rusher, and more."
Nice explanation, also actual completion rate can be misleading because it does not take into account throws that the QB should have made but did not. In CJB's case, I think his completion rate flatters him because he holds onto the ball a lot and gets pummeled in situations where Mullens would just throw the ball away to avoid a sack. Obviously the latter hurts your completion rate, but it also keeps you on your feet and off the injury list.
Good stuff. And not everything we've seen from CJB has been bad. He's had a good number of flashes.
I was at the Niners game in Landover MD (against what is now the WFT) when Hoyer got benched and CJB took over. He almost got us the win. One more first down in fact would likely have iced it for us (assuming Gould makes a kick he makes all the time). Garcon also went down in that game and ended up out for the rest of the season. I still wonder how much that hurt CJB's start to his time as a Niner?
A lot. So many young QBs have been ruined. I might be wrong on one of my stats above - the stat glossary was very vague on what it meant - but the point still stands - Beathard has not exactly gotten a fare shake.
yeah. I did not expect to see these numbers - to say the least.
Everyone is dead. Awesome.
sadly
Deebo should be back soon though!
one less dead person! I SEE THIS AS AN ABSOLUTE WIN
Everyone, Germ is not dead ... yet.
Oh you are alive!! The wonders of Internet right here!!
Good to see ya, Germ, how's the Bidet? I hear there are Goats up in Seattle who have contracted Covid, the culprit?
This is a really interesting article. And the statistical analysis, to me speaks to a key point of consideration: the distinctions between qualitative and quantitative analysis. I think it’s overly simplistic to say CJB’s receivers were bad and Jimmy’s were good. Or that BDN was luckier with the same cast (or something to that effect). I think part of the distinction is the QB’s “understanding” of his receivers. Not just getting the balls where they “should be”, and where they are “catchable”. We often hear Jimmy speaking of how he understands how his receivers are running their routes etc so his passing is “nuanced” (and this would be the qualitative difference that I use to “explain” the gulf in the difference in the rates of completion—Jimmy doesn’t try to pass “to” his receivers, but “for” them—this might also explain Jimmy’s reluctance to pass to Pettis). So while CJB can “make” the throws, he is not “passing” the ball with the same quality as Jimmy.
Anyway, that’s my speculation. It’s not as simple as Jimmy had better receivers. Jimmy contributes to making his receivers better. CJB hasn’t done that ... yet.
absolutely. Jimmy makes his WRs better which creates the illusion that they were already good. but its interesting to see that CJB's weaknesses might not be weaknesses. Garoppolo is a more advanced QB but I think beathard has something left.
Really good article to read Riqo, thanks. My aside about Pettis, maybe he runs his routes well, but is just a little “skatty” (or shaky) in how he’s finishing them off so Jimmy just doesn’t quite “understand” where he needs to put the ball for him. That pass in the Cards game, Shanny noted Pettis got bumped off his line early in the route. Another breakdown vid (johnnydels) said Pettis ran a slighty shallow angle in the deep route that left him short of where he “should” have been. So it’s like Jimmy’s passes to Pettis are basically hail mary’s—he puts them where they’re supposed to be and hopes Pettis gets there...
I think that sums it up well. CJB can make the throws on schedule but doesn’t set the game on fire.
There are lies, damn lies and statistics. CJB just doesn’t pass the eye test in a game.
I get why we have 3 QBs, if it hadn’t been a Covid year, I wonder if Kyle would have kept Broc to bring along and lose CJB as he couldn’t beat out Mullens.
That's what im saying though - i was a staunch supporter of cutting beathard prior to this season. IMO hes a bust, and a big one - but this made me rethink things. Shanahan obviously knows more about this offense than we do. And although Mullens has outplayed Beathard, Shanny referred to the two as equal during training camp. so maybe he sees what these stats also see in Beathard - that he isnt as bad as we think. Food for thought. I still think Mullens is getting traded once Jimmy is healthy - he hits FA after this season and if the 49ers can get a pick from a desperate team, theyll pull the trigger.
No one is going to trade for Mullens. There are a lot of Teams who may have similar backup QBs. Just remember Shanny's scheme and playcalling makes it a tad easier for the QBs to look competent, if you have a minimum skill-set
True but teams get desperate and injuries change things. In 6 weeks is the trade deadline - lets say Steelers are contending and Big Ben goes down - they won't play Rudolph or Hodges who were abysmal. They'll be interested.
Same scenario with Vikes and a couple others too
Teams inquired about Mullens the last 2 offseasons. I imagine that means they were prepared to trade for him at the right price.
I still don't get keeping 3 QBs, even this year. If your starter goes down, you can always get someone off the street to backup the 2nd stringer. If he goes down, your season is toast anyway.
i think its cuz Shanny offense is so complex. hard for a guy to come off street and learn.
I think last year we kept 3 as Jimmy was coming off an ACL. Made sense.
This year we're coming off a superbowl run, we're in the "window", so we've kept 3 to maximise chances of staying competitive if Jimmy went down for a period of time.
Again, makes sense.
I might be misunderstanding your analysis, but if I read it right you've fundamentally misunderstood the expected completion stat. Apologies if I've read your analysis wrong.
Expected completion is what you'd expect an average QB to complete based on the situations he's been in (taking into account distance, seperation, pressure etc.). It doesn't mean "what you'd expect the WR to catch given how well he threw the ball".
In other words, if I was the QB and attempted the same throws as CJB, then my expected QB would also be 68% - because it's not measuring how good my throws were, it's measuring how hard the situation was.
So Beathard's 68% expected completion stat does not mean he did his job well - it means he was put in situations where you'd expect an average QB to complete passes 68% of the time.
In Shanahan's offense, that number is always high - he puts his QBs in situations to succeed. So when you say "best in the league", that's misleading because it suggests CJB was achieving it. A better description would be "easiest throws in the league" - and yet he still wasn't succeeding.
The bit you are attributing to the receivers - CJB's 8% below average - is the CPOE stat, and is a pretty useful QB stat (and not a useful receiver stat).
This matches the eyeball test. Shanahan's scheme wasn't any worse with CJB on the field - the QB was.
Might be a different stat you are looking at. my stats are derived from NGS - they are defined as higher quality = higher xCOMP vs. lower quality throws = lower xCOMP.
https://www.nfl.com/news/next-gen-stats-intro-to-completion-probability-part-ii-0ap3000000967238
Try that ^. but from what I understand - a higher xCOMP is derived from the NGS completion probability stat and so that would mean whatever the xCOMP% is = the amount of high quality throws they made, regardless of completed. So the differential - that would be how the actual compared to what should have been completed (by the Qb and WR) - in this case, CJB being so high in that category means that he made a high percentage of quality throws (based on their NGS completion probability) - and the differential means that he got unlucky on those actually being completed
Yeah, you have to read carefully what the stat actually does. I'm very much in the "pro-stats" camp, but you have to be careful to understand what the stat is actually measuring (because the main reason we need next-gen stats in the first place is that last-gen stats are totally confused as to whether they are measuring player performance, coaching performance, team performance etc.).
[Or you can not to that. It's worked well for David Lombardi....]
yup. I read it (skimmed it) but the quality of situation rather than pass confused me. thanks!
"Expected Completion Percentage (xComp) gives an indication of the level of difficulty of a quarterback's throws. A high xComp indicates a quarterback's propensity to throw higher probability passes, while a low xComp tells us a quarterback has a tendency to take more chances."
Right, but "high probability pass" means a safe pass, not a good pass.
See the preceeding sentence to the one you quoted at the link below "Expected Completion Percentage (xComp) gives an indication of the level of difficulty of a quarterback's throws."
It's a measure of difficulty, not a measure of quality.
Yeah looks like you might be right. This is my first stuff with advanced stats so still learning 🤣 - but this also means a couple of articles from big outlets ive read are wrong too. Weird, but yeah you are right. Their explanation is super weird, at least to me. Edited (but the aggression and LCAD still stand.)
Not surprised even a little bit that major sites get it wrong.
And later from the same link you provided:
"Completion Percentage Above Expectation (xComp +/-) measures a quarterback's performance relative to the difficulty of their throws on an aggregate level."
The stat that CJB sucks at is the one that measure his quality. The one he is "best" at is the one that measures how easy his passes are (i.e. essentially the one that measures Shanahan's quality!)
Yeah, that sounds logical regarding Completion Percentage Above Expectation. If you are completing passes higher than the expected figure then you're probably good to very good.
I could be wrong. And even if I am - your point stands - Shanahan consistently makes his QBs make easy throws, which is why Mullens, Jimmy G, and Beathard all are similar in this department. But its interesting.
i think you might be right. NGS did a terrible job of explaining it. but the probability of completion stat makes me hesitate - "The probability of a pass completion, based on numerous factors such as receiver separation from the nearest defender, where the receiver is on the field, the separation the passer had at time of throw from the nearest pass rusher, and more."
this is confusing haha
Nice explanation, also actual completion rate can be misleading because it does not take into account throws that the QB should have made but did not. In CJB's case, I think his completion rate flatters him because he holds onto the ball a lot and gets pummeled in situations where Mullens would just throw the ball away to avoid a sack. Obviously the latter hurts your completion rate, but it also keeps you on your feet and off the injury list.
Good stuff. And not everything we've seen from CJB has been bad. He's had a good number of flashes.
I was at the Niners game in Landover MD (against what is now the WFT) when Hoyer got benched and CJB took over. He almost got us the win. One more first down in fact would likely have iced it for us (assuming Gould makes a kick he makes all the time). Garcon also went down in that game and ended up out for the rest of the season. I still wonder how much that hurt CJB's start to his time as a Niner?
A lot. So many young QBs have been ruined. I might be wrong on one of my stats above - the stat glossary was very vague on what it meant - but the point still stands - Beathard has not exactly gotten a fare shake.