Bryan Frye, owner and operator of the great site http://www.thegridfe.com, is back for another guest post. You can also view all of Bryan’s guest posts at Football Perspective at this link, and follow him on twitter @LaverneusDingle.

---

On Monday, I looked at which quarterbacks since 1960 helped or hurt their teams the most by taking or avoiding sacks. On Tuesday, I looked at the quarterbacks who gained the most or fewest yards through the air per attempt or dropback, and on Wednesday, we looked at completions relative to league average. Yesterday, the metric of the day was touchdown pass rate.

As promised, this article, “Dr. Safelove or: How I Learned to Start Worrying and Fear the Bomb,” centers on interceptions. The methodology here is no different from before: Figure out each player’s rate stats relative to the average of the rest of the league minus that player (LMP) that year and multiply it by his attempts to find the marginal total.

The caveat for this article is a big one: it is mathematically impossible for modern players to rank highly on a per play basis. In 1945, Sammy Baugh threw interceptions at a rate 7.4% lower than his peers. Because the league average today is so low (about 2.5%), a current quarterback would have to throw negative interceptions to match a -7.4% relative pick rate. Even if a quarterback threw 700 passes without an interception, the best he could possibly do is about -2.5%.

For single seasons, this makes a big difference. For entire careers, however, it isn’t as big of a deal. The higher volume of modern passers helps offset the superior rates of older quarterbacks. It’s not perfect, but hey … this is a data dump. ^[1]I thought about using standard deviations instead of just subtracting from the league average. This would be more mathematically honest, but it’s also beyond the scope of a simple data dump.

Single Season Stats

The first table presents the 1,563 passing seasons with at least 224 attempts and is sorted by marginal interceptions (lower is better). Read it thus: In 1947, Sammy Baugh threw 15 interceptions on 354 attempts, or 4.24% of his passes. The rest of the league threw picks on 8.73% of passes, so Baugh’s marginal rate was -4.49%. This gave him -16 marginal interceptions.

Like I said, the significant decrease in interceptions over the years makes it impossible for a recent player to rank highly. I’m not really worried about that, because I really want to highlight efficient seasons from bygone eras.

I’ve discussed Baugh ad nauseam in these posts (and for good reason). For now, I’ll just say: hey, he’s number one again.

Steve Bartkowski was decent quarterback on some exceedingly mediocre Falcons squads. He picked up a few Pro Bowl nods, but no one put him on the level of a Fouts or Montana (or even Ken Anderson) among his contemporaries. However, he did lead the league in touchdowns (1980), completion rate (1984), and interception rate (1983) once apiece. The passing environment in 1983 was such that a quarterback would be expected to throw about 19 picks on 432 attempts. Bartkowski threw 14 fewer than expectation – good for second best all time.

Sonny Jurgensen had the misfortune of playing on some pretty awful teams. However, despite playing for more than one team, in more than one system, and with a cast of uninspiring teammates, Jurgensen consistently produced on the field. He was a contemporary of Bart Starr and Johnny Unitas, but it was he who was considered the finest pure passer of his generation. In 1967, he led the league in both pass attempts and interception percentage. That combination resulted in 13 interceptions below expectation.

Roman Gabriel is all but forgotten by most modern fans (but certainly not by loyal FP readers), but he deserves much credit for his ability to avoid throwing the ball to the wrong team. No, Gabriel’s 3.3% interception rate can’t compare with Aaron Rodgers’ 1.6% or Tom Brady’s 2.0%, but for his era, he played it about as safe as a passer can play it. Four years apart, and with two different teams, Gabriel led the NFL in interception rate and saved his teams 13 picks.

The highest ranked season of recent vintage is Damon Huard’s magical one interception campaign in 2006. It was a complete fluke, as stats like these often are, but his 0.4% interception rate still ranks number one in history (among qualifying passers, obviously). ^[2]I think we can write this season off not only because of the rest of Huard’s career, but also because this performance came at the tail end of a half decade of efficient passing seasons from three … Continue reading

There are two notable passing seasons that are omitted by using a 224 attempt threshold. In 1936, Arnie Herber threw 13 interceptions on 173 passes, which is good for a 7.5% rate. That sounds pretty terrible at face value, but the rest of the league threw picks on 13.8% of their passes (it was a different age, folks). If he threw the ball to opponents at the same rate as his contemporaries, he would have tossed 24 interceptions.

Whizzer White’s 1938 offering is a different story entirely. On just 73 passes, he threw 19 interceptions. His rate of 15.9% was even bad for pre-WW2 passers. He gave the ball away at such a high rate that he produced twelve interceptions above expectation, despite his low volume. Good thing he could fall back on being a Supreme Court Justice. ^[3]Also, White led the league in rushing yards and yards from scrimmage in 1938. It was a different time, folks.

Career Stats

This table shows the 222 careers with 1,000 or more pass attempts. You probably know how to read these things by now.

There are effectively three tiers at the top of this list: the Roman Gabriel tier, the Fran Tarkenton tier, and the Joe Montana and Tom Brady tier.

In his sixteen years with the Rams and Eagles, Gabriel never threw interceptions at a rate worse than average. ^[4]If you check out his PFR page, you’ll see his lowest INT%+ score was 108 in 1972. Over the course of his 4,498 attempt career, he saved his teams 94 interceptions. ^[5]He balanced it out by throwing touchdowns at a below average rate, but we’re focusing on positives here. Think about it like this: the formula for ANY/A values interceptions at -45 yards. That means he saved 4,230 adjusted net yards by playing it safe.

Tarkenton’s career began a year before and ended a year after Gabriel’s. He ranks a distant to his rival in marginal interceptions; however, unlike Gabriel, he didn’t sacrifice positive plays while keeping his turnovers to a minimum.

Montana and Brady are often lumped together by fan and analysts alike (I can’t imagine why). Here, they sit side by side in their own tier, far below Tarkenton, but far above Baugh. Both quarterbacks thrived in systems that emphasized their intelligence and accuracy, which led to high completion and low interception rates for both. However, let’s not forget that the two also gained yards and threw touchdowns at a higher rate than their peers as well.

I want to avoid being negative today, so I’ll wrap a negative in a positive. Ken Stabler, threw 37 interceptions more than expectation over the course of his career. However, the rest of his rate stats compare favorably to almost any passer in history. This is pure speculation, but his interceptions may be the primary reason he doesn’t currently have a bust in Canton. ^[6]I think it’s important to mention that interceptions weren’t as harmful in Stabler’s era as they are now. The dataset on which ANY/A is based comes from an era of increased offensive … Continue reading

Another article, another shout out to Cecil Isbell and Ed Danowski. They don’t reach the thousand attempt threshold, but they rank very highly in marginal interceptions. In 818 attempts, Isbell threw 28 fewer picks than you’d expect. In 637 attempts, Danowski was 27 picks below expectation. Those numbers would rank 29th and 30th all time.

That’s it for me, friends. Thanks for helping me steward the site in Chase’s absence. Leave your thoughts in the comments.