I think you should should give Maddux Madsen a shot. Yeah, you. You have no influence on depth charts, try as you might on the Twitter battlefields, but you should give Maddux Madsen a shot.

This isn’t about Malachi Nelson, this isn’t about Taylen Green (though we’ll talk about him too), this is about you. Give the kid a shot, will ya?

See, I get it. I was there, too. QB swaps and angst and frustration and stupid losses and 2023 did a number on fans, MW chip notwithstanding. I’ll get reminded we went 8-6 and that seems to come as a shock to me every time.

SOME of you, not of all you, but some deep down blame Mr. Madsen for such woes. Now that Andy is gone, MM gets all the angst. Believe me, I’ve heard it. He kinda like threw a bunch of almost-picks against Fresno and wasted Taylen Green and all that jazz… Listen, I won’t try to unpack all of our bad feelings, but just have an open mind, yeah?

Stats, Advanced

My dad is fond of broadcasts that note obscure stats. Did you know Ashton Jeanty will be playing in his first game ever to be on a Tuesday and a full moon? Neat.

Now, those aren’t really stats. Anecdotes maybe, but you know what is useful? Context. The big term for today is EPA: expected points added.

Let’s say quarterback A completes a pass for 17 yards. Pretty good, right? Now let’s say quarterback A’s team is losing by 30 and the pass was a screen pass thrown on 4th and 30. Not so good anymore, not even a first down. Then let’s say quarterback B completes a 17 yard pass on 4th and 12 to seal victory a over a highly ranked opponent. That’s amazing! Problem is, when we look at the box score…they have the same stats.

A popular solution to contextless, Godless box scores is EPA. In EPA land, every play is worth points or fractions of points.

An oversimplified explanation can be found on the football field. You have a 50% chance of making a field goal? Your expected points on that kick is half a field goal, 1.5 points. If you make it, you’ll have 3 points, 1.5 more than expected. Therefore, your EPA is 1.5.

One more example: Team A has a first and goal on the one yard line. They throw a pass and team B intercepts, running it back for a TD. Team B had an expected points of -7 on the play, since they were about to be scored on, but they ended up scoring 7. Therefore, they have an EPA of 14 on that play. It gets more complicated on non-scoring plays, but a little bit of math estimates the process for us.

EPA (No, not that EPA)

In November, I posted a thread discussing EPA in the QB room. Given the QB controversy, it opened to mixed reviews.

The cliff’s notes is that EPA was much higher for Maddux Madsen than Taylen Green, specifically passing EPA, but total EPA as well. On the flip side, George Holani and Ashton Jeanty were more efficient (i.e had a higher EPA) with TG on the field.

My goal in posting was not to persuade the coaching staff to reconsider their lineup decisions. They have information that far exceeds what I was able to pull from collegefootballdata.com.

See, I was interested in persuading the masses. Not even that Maddux Madsen should start over Taylen Green, simply that Maddux Madsen was playing well.

Okay, but is he good?

Now that we’ve played a whole season, it’s time to revisit that. Shown below is the EPA of every QB since the 2014 season:

Before we overreact, say it with me: SAMPLE SIZE. I’m not here to tell you Mr. Madsen will exceed the likes of Grant Hedrick and Brett Rypien. I am here to tell you that Maddux played well, at times really well (Wyoming game) and that you can let yourself believe a bit.

A measured take would be not that Maddux Madsen is better than Grand Hedrick or Brett Rypien, but that he has earned his snaps with efficient play. In a way this table shouldn’t be surprising. EPA is expected points added, which rewards players for doing better than expected.

Maddux Madsen was generally deployed during 2nd/3rd and long, 2 minute drills, and when the Broncos were facing a deficit; all difficult situations with low expected points. But that’s context that matters. Most QBs need throws to get into rhythm, so getting almost exclusively put in bad situations and playing efficiently is impressive.

If this is your first time learning about EPA, I’m guessing I haven’t convinced you. You watched Maddux Madsen with your own eyes, a shiny color coded table probably won’t change your mind.

So let’s look under the hood on these EPA stats (And more shiny color-coded tables.) Knowing that attention spans are limited these days, I’m gonna go with two pieces of evidence: conversion rates and deficit play.

Conversion Rates (Not the Metric System)

If EPA is all about making good plays when it matters the most, what plays matter the most? You might disagree on that choice, but 3rd/4th downs feel like valid candidates. Converting those is good, IMO.

Below I have a table of 3rd/4th down conversion rate by distance. I cut our player list to just Maddux and Taylen to spare your eyes some extra work.

A few things to point out: Taylen, better at short yardage situations, who would’ve thought? Also somewhat better on the 10+ yard plays, which is maybe more surprising, but neither .3 or .27 are particularly high conversion rates.

The interesting place to note is in that 3rd & 7-10 range. Maddux converted 50% from that spot on 16 plays. It’s an unusually high clip, and does something important: extends drives.

An underrated benefit of this is better defense. A holistic view of defense tells us that a team’s offensive play influences their defensive play. An efficient offense that converts on third down and extends drives will give their defense more rest, helping to improve their play.

Taylen was a supremely explosive player (shoot I mean watch his highlights from last night), but the reality is that playstyle sometimes gave spectacular 80 yard TD runs, and sometimes 3 & outs. I’ll miss those plays, but there’s an argument that slow and steady wins the race.

Deficit Spending

Another way to look at contextually important play is how we play when we’re losing. If we remember games like UCF, Memphis, San Jose St, this is where Maddux shined. But do the stats back that up? Take a look:

I highlighted EPA because, well, that’s kinda the point of my article, but take a look at the rest of the numbers. You’d love to get rid of Mad Dog’s picks, but it’s hard to complain about that passing/rushing combo.

While I’d prefer to not face many deficits in 2024, that’s good play under pressure. EPA rewards that kind of thing, which again explains why EPA loves him so much.

“Turnover Machine”

To present one last argument for Mr. Madsen, it’s time to address the “turnover machine” allegations. Many members across Bronco nation have expressed concern that starting Maddux Madsen is risky because of turnovers. Some have decried his height, others his decision making.

Broncos fans are acutely aware of this from one game in particular: our trip to Fresno last year. In that game Maddux Madsen played the majority of the snaps, throwing for 258 yards on 22/42 passing, with 1 TD and an INT. He also had 9 carries for 61 yards, including a 27 yard TD run.

On it’s head, that statline sounds better than the game felt, and that might be a good point. Many will be quick to point out that Maddux Madsen “almost” had 3 more picks. Let’s talk about that.

On one hand, “almost picks” or “should’ve been picks” don’t count in the game. They don’t really count in EPA either. However, they do count when evaluating QBs. Unless Maddux has some kind of special spiral that makes it hard for the DBs to catch, they will get interceptions eventually if that kind of play continues. Which side am I on? Context, always context.

Special Stats

PFF has a useful tool for measuring these kinds of things, called “Turnover Worthy Plays”. For example: it’s probably fair to say the QB is not to blame for plays like this:

Thanks to Nathan Carroll on Twitter, I was able to procure some useful PFF stats, including Turnover Worthy Plays (TWP):

In English, Maddux and Taylen had essentially the same rate of turnover worthy plays, about one every 20 plays. The level to which you think that’s good probably relies on what you thought of Taylen as a QB. The reality is that Taylen was a pretty (and sometimes very) accurate quarterback.

His interceptions weren’t often the result of throw accuracy or ability so much as timing and decision making. TG’s average TTT (Time to throw, another PFF stat, shoutout Nate) was 2.85, a fairly slow time in a fast game of football. A guy who was open at first may not be open later, and a 6’6 frame necessitates a longer throw, working against Taylen.

When looking at MM, his interceptions were due to his height and decision making. Madsen made quick decisions with the ball, and when he didn’t, those lead to interceptions. The infamous pick at Fresno was an example of that.

His TTT was 2.58, which doesn’t sound like a lot, but is much faster than Taylen and is elite at the college level. In this case, Maddux’s stature (and thus throwing mechanism) works to his benefit in getting the ball away from situations that might be picks for Taylen.

Even More Context

You still may not be convinced, and I’m getting tired of persuadin’. I have one last argument for you. Was Maddux bad, or did he just have a bad game against a good team?

Now, if we want to be really good, we’ll need to be able to throw against teams better than Fresno, but I think we can remember this compelling point: he was a freshman (and getting half the practice reps. So, technically, *two* compelling points.)

Truth be told, Jeff Tedford’s teams have played us well the last few matchups. Taylen ran all over a depleted Bulldog team in his second career start, but struggled mightily in his MW championship game outing.

(Taylen Green against Fresno)

A lot of people (myself included) felt like we abandoned the run in that ’22 chip, but it would seem the story was more about Fresno stacking the box and forcing Taylen to pass, which worked great for Fresno. That probably also explains why Maddux Madsen came in and threw as much as he did in 2023.

Fresno led the conference in interceptions (16) last year, and 4 of Maddux Madsen’s 8 Turnover-worthy plays came against them, which would indicate a struggle against them more than a general decision-making problem.

Again, the measured take here is not that, “Maddux Madsen doesn’t make mistakes”, just that “turnover-prone” is hasty. We quickly forget he led two straight scoring drives (and probably would’ve led a game-winner) had it not been for a painful kickoff return and a lot of bad tackling. For a rough start, he deserves credit for that game too.

One Last Thing

Maybe you don’t believe in math, or stats, or maybe just my stats. That’s fine. Maybe things go south for Maddux and that’s life, that’s football. But if that’s you I’m talking about, give him a shot.

Undersized, overlooked, this kid is all heart. Hyperbole, yes, but I mean it. You’re telling me you didn’t feel something when he bowled over a LB on 4th & 2 to keep the drive alive against UCF? When he connected with Cobbs for the go-ahead score?

A lot of you want to talk about ceiling. Yeah, I think Malachi Nelson has a high ceiling, but you think what Maddux has just grows on trees? Took a stalling offense to two straight TDs vs. Memphis, lead two straight scoring drives against a Fresno team that seemingly had his number all game. Not everyone does that, not everyone can do that.

Maddux Madsen has a high ceiling too. Better QBs have crumbled under the pressure of a late drive, but I swear that kid relishes it. That kind of moxie is what a playoff team needs. All I know, is that there will come a time when we need to win a game throwing the ball. Jeanty will take care of most, but if it comes down to it, Mad Dog will answer the bell.

Look, I could be wrong about all of this. Maybe last year was a fluke, or maybe he doesn’t really improve all that much this year. Maybe he only starts one game. But what about this: what if he’s good? What if he’s really good? What if he gives us something no one else can? Look, it’s week 1. It’s free to believe.

Give ‘im a shot.

Author