Originally Posted by :
Patrick Mahomes is the rarest of athletes, with so much physical talent that expectations stretch beyond what should be realistic and so much confidence that his level of play goes even further. He is a revelation.
The experience of watching his career with the Chiefs from the ground up has basically been a weird tug of war between what a logical mind thinks possible and what open eyes see as reality.
So, here’s a moment won by open eyes:
Patrick Mahomes — 23 years old, in his first season as a starter and just his fourth as a full-time football player — will be the NFL’s MVP if he closes even nearly as well as he’s begun.
We can do this statistically. We can do this schematically. We can do this emotionally. We can even do this via narrative.
The conclusions are the same, and even allowing for the obvious caveat that the last four games of the season will carry disproportionate importance in the minds of voters, he has been the best player on the best offense and most important position in American sports.
“I don’t want to disrespect him with this, but I see a lot of my young self in him,” Packers quarterback Aaron Rodgers said.
“That’s the future of the league,” Patriots quarterback Tom Brady said.
“We look at him as a young Brett Favre but just more athletic,” Ravens cornerback Jimmy Smith said.
The MVP is among the most subjective awards in any sport. That might be particularly true in football, which is harder to quantify than baseball, for instance. What’s most valuable to you might be different than what’s most valuable to me.
But at least at the moment, Mahomes’ case borders on overwhelming.
He leads the NFL in touchdown passes, touchdown percentage and total quarterback rating. He is second in yards and passer rating. The NFL has made itself into a league ruled by offense. The Chiefs have scored more points than anyone else, and Mahomes is the single biggest reason.
He has thrown as many touchdowns as Rodgers and Brady combined, and as many as the man he replaced threw in the last two seasons combined. He is on pace to tie Peyton Manning’s all-time record of 55 touchdown passes in a season. There is a logical case to be made that Mahomes is having the best season by a quarterback in NFL history.
Drew Brees is the closest thing to Mahomes’ statistical equal — first in passer rating and completion percentage, with 30 touchdowns (third) and just three interceptions. He is a Super Bowl champion and first-ballot Hall of Famer who might benefit from a sentimental vote. But he also threw for just 127 yards, one touchdown and an interception in an ugly and nationally televised 13-10 loss at Dallas. That will be tough for voters to forget.
This sounds made up, but Mahomes’ worst game of the season was either the time he won at Denver on Monday night with a left-handed pass, threw for more than 300 yards in a blowout win over Jacksonville, became the youngest quarterback to throw for 350 yards and four touchdowns against Bill Belichick, or that time he threw for 478 yards and six touchdowns against the Rams.
One more time: Those are his worst games.
If you are the type to take the “V” in MVP literally, then Mahomes’ case is conveniently and uniquely clear.
A year ago, with a quarterback who has shown himself to be somewhere between average and good, the Chiefs finished fifth in yards and sixth in points. Alex Smith had an MVP-worthy first five games, then mostly struggled thereafter, including zero points after halftime in an embarrassing playoff loss.
Mahomes now leads the same offense. Same head coach, mostly the same teammates. Tyreek Hill and Travis Kelce are again the top pass catchers. Mitchell Schwartz and Eric Fisher are again the starting tackles. Injuries have again required reshuffling the interior of the line. Sammy Watkins (when healthy) is an upgrade as the No. 2 receiver, but the biggest difference other than Mahomes is that Kareem Hunt has just been released.
Professional football includes too many moving parts for a controlled experiment.
This is the closest we will ever see:
The Chiefs are averaging 11 more points and 62 more yards. They have already scored 14 more touchdowns. That’s because of Mahomes.
A year ago, with a system based on timing and scheme and matchups, the Chiefs scored more than 35 points twice and averaged 26 points. This year, with a system that includes all of that but also features a supernaturally talented quarterback, they are averaging 37 and haven’t scored less than 26.
Their defense has actually been worse, but the team has gone from a 10-6 division winner to 10-2 and among the betting favorites to win the Super Bowl.
In other words: Mahomes has pushed this team from a representative division winner to a real-life Super Bowl contender, despite the defense being statistically worse.
If that’s not an MVP, what is?
Mahomes would be the first MVP in Chiefs history, and the youngest since Dan Marino in 1984 — also 23, also in his second year in the league. And he would have earned it. The Chiefs’ next three games are against three top 10 defenses — Ravens (first), Chargers (seventh) and Seahawks (ninth).
That’s a brutal finish, even without this now being the longest season Mahomes has played since high school. He will be tested in new ways, by some of the best minds, bodies and defenses in the league.
The Ravens are a particular challenge. They have the best pass defense in the league (5.1 yards per attempt). In four games against Atlanta’s Matt Ryan, New Orleans’ Drew Brees and Pittsburgh’s Ben Roethlisberger (twice), they gave up an average of 212 passing yards.
The Ravens blitzed Ryan and Brees more often than not, according to Pro Football Focus, and there was a time that would’ve been a major worry for the Chiefs.
Mahomes came to them with unique talent, but the coaching staff saw his blitz recognition as one of his biggest needs for improvement. This week, Reid singled out that specific trait when asked where Mahomes has progressed faster than expected. He didn’t add this part, but the Chiefs saw this almost immediately, when Mahomes shredded the Steelers’ “Blitzburgh” defense for 126 yards and two touchdowns on 12 blitzes.
“He was better than we thought,” Steelers linebacker Bud Dupree said after that game. “I feel like his intelligence was better than we thought. We thought since he was a young quarterback we could trick him.”
The mistake Dupree described there was planning for what a logical mind would think possible.
We are now 12 games in, with enough evidence to believe in a reality greater than anyone could have imagined.
That's an interesting video. The comments are also interesting. I think a lot of the negative reaction from fans to the clusters that PFF described comes from PFF misdescribing the clusters as "tiers" and the implication that they put the QB in tiers using mathematics. In fact, that's not really what they did.
In the video, the analysts say they used principal component analysis (PCA) to cluster the QBs. To understand what this does and does not allow you to do, let's talk about when and when we can't put geometric points in order.
Points on a line are well-ordered. That is, if you took two distinct points on a line, say -3 and 14 you can always determine whether one is less than the other. Points on a plane are not well-ordered. Suppose you took three points on a plane, say (3, 2), (2, 3) and (4, 4), then you wouldn't necessarily be able to say that the point (3,2) is less than or equal to the point (2, 3), even though you may be comfortable in saying that the point (4, 4) is better than both of them. The reason why this is relevant is that PCA summarizes a large set of numbers into a set of component scores that can then be used for purposes of clustering. In practice, to visualize the clusters, the first two components are used. You can see such a visualization on this webpage, which is from PFF and describes their clustering of the 2017 NFL QBs. See that figure toward the bottom, near the Conclusions. https://www.profootballfocus.com/new...qb-performance
If you look at that figure, you will see that there is not any obvious way to look at "cluster 2" and "cluster 3" and decide that one is better than the other. If you read the PFF article with the understanding that PCA does not produce an ordering and that one would need some external information to provide such an ordering, you will see that in fact the PFF article even acknowledges that a so-called Cluster A is not necessarily better or worse than a so-called Cluster B.
I say all this to point out that it's incorrect for PFF to use the terminology "tiers", which implies an ordering, and for PFF to pretend that they derived those "tiers" mathematically. Instead, what PFF did was use mathematics to form clusters. Then they looked at who was in the clusters and their own external judgments about who is top-tier, middle-tier, bottom-tier to label the clusters with numbers as though they are well-ordered. They'd be better off just labeling the clusters with letters and pointing out the cluster that has the most "elite" QBs and not bother trying to rank all of the clusters they get, because in fact it's not necessarily the case that the clusters can be put in a strict order. In particular, the cluster of mediocre QB who are game managers is neither better nor worse than the cluster of mediocre QB who plays aggressively but also make many mistakes.
Of course, we have an elite QB and he's definitely top-tier! I'm just explaining why there is a lot of negativity in the comments toward that video, which is that the fans realize that it's not sensible for PFF to try to act like PFF knows who's 2nd vs 3rd tier. [Reply]
Leagues #1 defense had 2 shots to slam the door on Mahomes.
MVP overcomes teammate shitting his pants causing 4TH n long with a false start with one of the most insane scrambles/throw on the run Ive ever seen, and then overcomes an unfathomably stupid playcall by Andy on 3rd n 5 to score on 4th down and tie the game.
Originally Posted by DanT:
That's an interesting video. The comments are also interesting. I think a lot of the negative reaction from fans to the clusters that PFF described comes from PFF misdescribing the clusters as "tiers" and the implication that they put the QB in tiers using mathematics. In fact, that's not really what they did.
In the video, the analysts say they used principal component analysis (PCA) to cluster the QBs. To understand what this does and does not allow you to do, let's talk about when and when we can't put geometric points in order.
Points on a line are well-ordered. That is, if you took two distinct points on a line, say -3 and 14 you can always determine whether one is less than the other. Points on a plane are not well-ordered. Suppose you took three points on a plane, say (3, 2), (2, 3) and (4, 4), then you wouldn't necessarily be able to say that the point (3,2) is less than or equal to the point (2, 3), even though you may be comfortable in saying that the point (4, 4) is better than both of them. The reason why this is relevant is that PCA summarizes a large set of numbers into a set of component scores that can then be used for purposes of clustering. In practice, to visualize the clusters, the first two components are used. You can see such a visualization on this webpage, which is from PFF and describes their clustering of the 2017 NFL QBs. See that figure toward the bottom, near the Conclusions. https://www.profootballfocus.com/new...qb-performance
If you look at that figure, you will see that there is not any obvious way to look at "cluster 2" and "cluster 3" and decide that one is better than the other. If you read the PFF article with the understanding that PCA does not produce an ordering and that one would need some external information to provide such an ordering, you will see that in fact the PFF article even acknowledges that a so-called Cluster A is not necessarily better or worse than a so-called Cluster B.
I say all this to point out that it's incorrect for PFF to use the terminology "tiers", which implies an ordering, and for PFF to pretend that they derived those "tiers" mathematically. Instead, what PFF did was use mathematics to form clusters. Then they looked at who was in the clusters and their own external judgments about who is top-tier, middle-tier, bottom-tier to label the clusters with numbers as though they are well-ordered. They'd be better off just labeling the clusters with letters and pointing out the cluster that has the most "elite" QBs and not bother trying to rank all of the clusters they get, because in fact it's not necessarily the case that the clusters can be put in a strict order. In particular, the cluster of mediocre QB who are game managers is neither better nor worse than the cluster of mediocre QB who plays aggressively but also make many mistakes.
Of course, we have an elite QB and he's definitely top-tier! I'm just explaining why there is a lot of negativity in the comments toward that video, which is that the fans realize that it's not sensible for PFF to try to act like PFF knows who's 2nd vs 3rd tier.
Just stop. I understand PCA and orthogonality. It's a much more complete way to define certain exposures. But with Mahomes - the eye test wins. He's a god amongst men. Go back to your econometrics class, nerd. [Reply]
