A small bone to pick with FiveThirtyEight

Lord knows, I love stats. Love what you can learn from them. Love just playing around with them to see what turns up. And what FiveThirtyEight.com’s Neil Paine does with stats in his revisionist piece about the Greatest Show on Turf — the 1999 Rams offense — is terrific. By all means read it, if you haven’t already.

My only quibble is Paine’s overinflation of Kurt Warner’s ’99 season. “Warner ended up completing 65.1 percent of his passes,” he writes, “which at the time was the third-best single-season completion percentage by any quarterback ever.11” Third-best ever. Wow. That one caught me by surprise. Then I chased down the footnote and found out he was talking about only “quarterbacks with 450 attempts.”

I’m not sure what, in Paine’s mind, is so magical about 450 attempts — other than that it allows him to say Warner’s completion rate was “the third-best . . . ever.” After all, 450 attempts are a lot of attempts. Only three NFL quarterbacks had that many in a season before 1978, when the schedule was increased to 16 games and rule changes turned pro football into the passer’s paradise we have today. (Note: Five more had 450-plus in the bombs-away AFL.)

But that’s a minor point because, the rules being what they were, almost no quarterback back then was going to complete 65.1 percent of his passes — unless it was the Redskins’ Sammy Baugh hitting 70.3 in the talent-starved war year of 1945. Show me a QB in those days who connected on 65.1 percent, and I’ll show you an extraterrestrial.

The larger point is: Why 450 attempts? Steve Young threw 447 passes in 11 starts for the ’95 49ers and completed 66.9 percent (1.8 percent more than Warner). We’re just going to leave him out? Then there’s Joe Montana, who threw 386 passes in 13 starts for the ’89 Niners and completed 70.2 percent (5.1 percent more than Warner). We’re going to ignore that season, too, even though Joe set a record that year (since broken) with a 112.4 rating?

What I’m objecting to is the arbitrariness of “450 attempts,” which serves no real purpose except to make Warner’s season look better. And here’s the thing: Neither he nor the story of the ’99 Rams offense needs any ginning up. His numbers are perfectly capable of standing on their own, without any creative massaging. It was, by any measure, a fabulous year, among the greatest of all time. For Paine create this imaginary 450 Attempts World — in which Warner has “the third-best single-season completion percentage by any quarterback ever” — is just plain silly.

To qualify for the passing title, a QB needs to throw 224 passes (14 per scheduled game). If you make that your threshold, Warner had the 17th-best completion rate ever. The Top 5:

HIGHEST SINGLE-SEASON COMPLETION RATES THROUGH 1999

Year Quarterback, Team Att Comp Pct
1982 Ken Anderson, Bengals 309 218 70.6
1994 Steve Young, 49ers 461 324 70.3
1989 Joe Montana, 49ers 386 271 70.2
1993 Troy Aikman, Cowboys 392 271 69.1
1993 Steve Young, 49ers 462 314 68.0

Minimum: 224 passes.

(Note: The schedule was only nine games in ’82 because of a player strike.)

Again, Warner had a sensational season, especially when you consider his 41 touchdown passes, 109.2 rating and Disneyesque backstory as a former Arena Leaguer. But making the cutoff 224 attempts, the league standard, instead of 450 tones down the idolatry a little — which is what statistical research is supposed to do.

Of course, 17th doesn’t sound nearly as good as “third-best . . . ever.” But what are you gonna do? It’s one thing to ignore Frank Filchock’s 111.6 passer rating for the 1939 Redskins because it isn’t “modern” — even though, coupled with his 413 rushing yards (ninth in the league), it was one of the most amazing years in NFL history. But when you disqualify seasons by recent Hall of Famers like Steve Young, Joe Montana and Troy Aikman because they fall short of some arbitrary minimum (450 attempts – and not a pass less!), that’s when I’m going to pipe up.

Source: pro-football-reference.com