New TAYP Quick Update

If you’re familiar with my work, you know I created the metric New TAY/P to measure quarterback play, using only inputs widely available on sites like Pro Football Reference, NFL.com, ESPN, Sporting Charts, etc.1 I wouldn’t put it on par with other advanced stats, such as Total QBR, DVOA, or EPA/P, which are the cream of the quarterback efficiency stat crop. However, it does have the advantage of being completely transparent and easy to calculate on a napkin, in case you’re the sort who likes to calculate things on napkins.

The formula:

New Total Adjusted Yards (NTAY) = [Yards + Touchdowns*11 – Interceptions*45 – Fumbles*25 + First Downs*9] / Plays, where

Yards = passing air yards + passing YAC/2 + rush yards – sack yards – yards lost on kneels

Touchdowns = pass touchdowns + rush touchdowns

First Downs = pass first downs + rush first downs

Plays = pass attempts + sacks + rush attempts – spikes – kneels

Simple stuff.

In case you are unfamiliar with some of the stats:

NTAY/P+ is the New Total Adjusted Yards per Play Index Score, which I derived using the method described in this post. It is similar to Pro Football Reference’s Index Scores, but it uses a weighted standard deviation instead of the standard deviation of just qualifying passers.

VAL (value) and VAL/P (value per play) are measurements of a player’s performance relative to that of the rest of the league over a three year span, with the given year in the middle.2 If a quarterback’s NTAY/P is 7.50, and the league average is 6.00, his VAL/P is 1.50 (7.50 minus 6.00). His VAL is then his per play value multiplied by his total plays. So if Mr. Hypothetical ran 605 plays, his VAL would be 908 (1.50 times 605, rounded).

REP is value above replacement. For the sake of easy, frankly, I chose 75% of the three-year league average NTAY/P as the baseline for replacement-level play. REP is then calculated the same way VAL is. This isn’t rocket science. It’s barely math.

Because the passing environment in 1992 was much different from that in 2016, I adjusted for workload based on era. When you see “AdjVal” or “AdjRep,” you’re seeing the standard VAL and REP metrics modified based on this workload adjustment.

I think I’ve explained all the nebulous terms. If I haven’t, feel free to comment or send a tweet my way.

I don’t have a ton of time to get into the nitty gritty, as it were, so I’ll provide a quick report that I think does a fine job condensing a lot of information into easily digestible charts.

The report consists of five pages:

  1. Career Stats
  2. Single Season NTAYP+
  3. Career NTAYP+
  4. Single Season Adjusted Value Over Average
  5. Career Adjusted Value Over Average

The default view shows every qualifying quarterback at once (300 plays in a season or 900 plays in a career, since 1992). However, you can click a name in the check box to view a single player, or you can control+click several names to look at a specific group of players. This is useful in case you want to compare Peyton Manning and Tom Brady, Brett Favre and Aaron Rodgers, Eli Manning and Tony Romo, Russell Wilson and Andrew Luck, or, of course, Colin Kaepernick and Josh McCown.

If you haven’t fallen asleep yet, take a look at the charts, and draw your own conclusions. I suggest clicking on the bottom right and going full screen.

  1. New TAYP is just a revamp of Total Adjusted Yards per Play. The difference is that the new version only gives quarterbacks half credit for yards gained after the catch. You can argue both for and against diminishing a quarterback’s credit for YAC, but one thing that’s certain is that 50% is not an exact number derived from laborious research. It’s a proxy for a future number and a launchpad for discussion.
  2. Except for 1992 and 2016. 1992 doesn’t have a preceding year in the data, and 2016 doesn’t have a succeeding year.
  • Adam

    These visualizations do a great job of showing how far Brady, Brees, and especially Manning are ahead of their contemporaries.

    • I like to look at it in tiers. If you select all and then control+click to remove the big three, then Young, Rodgers, Favre, Ben, Rivers, Ryan, Romo, and McNair, really stand out. Then you add the big three back in, and you see just how far ahead of the pack they are.

      Take away all those guys, and you see some more separation within the rest. Marino, Aikman, Green, Garcia, and Warner form a little group, while Brunell, Palmer and McNabb stand out as higher volume guys with similar value. Then you look at the bottom and remember that time Kerry Collins and Trent Dilfer squared off in the Super Bowl.

      • Adam

        That’s a really cool way of looking at this.

  • Four Touchdowns

    Hey, I didn’t realize you had your own football blog!

    What made you create this instead of rolling with ANY/A?

    • I think incorporating rushing, first downs, and air/YAC is important.

      • Four Touchdowns

        So so I. I especially like how you handle YAC.

      • Four Touchdowns

        Since you’re factoring in YAC to yardage, do you factor it into first downs? (first downs gained by YAC)

        • No. Unfortunately, I don’t have access to information like that, aside from my collection of game books. But game books require parsing often indecipherable text.

  • Four Touchdowns

    I know this is arbitrary, but the one smart thing ESPN did when they developed QBR was to make it a 0-100 scale.

    While I know that may not be feasible with these rate stats, I think of PFR’s ANY/A+ and Passer Rating + stats as a good way to “translate” those stats for the common fan. I just think that fans are so conditioned to passer rating that any new efficiency metric will have to at least look similar.

    • The NTAYP+ stat is the same as PFR’s index scores but with New TAYP instead of ANYA.