Introducing Tackles Per Opportunity
In the dark ages of fantasy football, we often gauged an individual defensive player’s value through his raw statistics. The logic was simple: Player A had 150 total tackles compared 125 total tackles for Player B. Thus, Player A is the better fantasy option.
A little under a decade ago, player evaluation experienced a quantum leap when we started to factor in the impact of home stat crew bias. We recently took another significant step forward when analysts began utilizing advanced metrics.
Here at PFF, we developed one such metric in tackle frequency. This stat allowed us to compare apples to apples by reducing tackle production to a per snap rate. While many in the industry (including yours truly) have extensively used tackle frequency over the past several years, this rate stat has its drawbacks. Today, I’m introducing a new statistic that will render tackle frequency archaic – Tackles per Opportunity, or tPOP.
At this point, the problems with using raw statistics have been fairly well documented, and I’m not going to rehash those points here. In fact, this general argument was the impetus for the creation of tackle frequency. With snap numbers finally available for all players, we could simply divide this data by total tackles to calculate tackle frequency. Finally we had a way to measure individual player fantasy value in a way that put all defensive players on equal footing.
Or so we thought.
Let’s take last season as an example. Among linebackers who played at least 500 snaps, Lavonte David (15.01%) and Bobby Wagner (14.79%) finished Nos. 2 and 3 respectively in tackle frequency. While both players are very valuable, convention wisdom would argue that David’s tackle frequency suggests he’s slightly more valuable.
Unfortunately, there’s a fairly significant blind spot in this logic: Tackle frequency assumes a player can make a tackle on all of his snaps. This myopic perspective fails to account for incomplete passes. Simply put, a player can’t make a tackle on plays with an incompletion. Tackle frequency completely ignores this fact.
In taking all snaps played (including playoffs) sans penalty snaps, David totaled 886 to 825 for Wagner. David was on the field for 114 incompletions, while Wagner had an incompletion thrown on 137 of his snaps. If we subtract incompletions from snaps, we get what I call a “tackle opportunity.” In this example, David finished with 772 tackle opportunities and Wagner had 688 tackle opportunities last season.
Now, we calculate tPOP in a similar manner to tackle frequency. Just replace raw snaps with tackle opportunities, and there you have it. So, David’s tPOP comes in at 17.23 percent, which is a half-percent lower than Wagner’s 17.73 percent. In fact, David also comes in just behind DeAndre Levy’s 17.25 percent, making him the No. 4 tackler on a per play basis.
Does this example significantly change what we think of David in terms of IDP value? Absolutely not. However, we now have a more accurate view of his tackle production from last season.
The key here is predictability. In looking at the numbers over the last eight seasons, I ran a regression analysis for a normalized version of tPOP. Here are the results:
An r-squared value on 0.97 indicates an extremely high correlation between tPOP and tackle production. Keep in mind that this is normalized data, but even the regular tPOP comes in at an r-squared 0.70.
It should also be noted that in these calculations, I’ve used PFF’s tackle numbers. Long-time readers will know that our game charters independently record tackles. While these aren’t the numbers you’re going to see in the box scores, they provide a more accurate view of player production. Since our numbers are coming from one crew, we don’t have deal with the problems of home stat crew bias.
While tPOP is another step forward in IDP analysis, it’s not a perfect metric. For starters, this stat has a built-in assumption that a player can make a tackle on any given play. Football obviously isn’t that simple. However, in looking at the numbers over the last eight seasons, certain players do make tackles more frequently than others and do so on a fairly consistent basis.
Luke Kuechly has ranked among the Top 5 players in tPOP in each of his three seasons, with an average tPOP of 19.13 percent over that span. On the other end of the spectrum, a player like Sean Weatherspoon (when healthy) has also posted tPOP numbers that consistently fall outside of the top 30. His career tPOP of 13.91 percent indicates that we should temper our expectations for Weatherspoon this season in Arizona.
Let’s take a look at some of the tPOP numbers from last season. Displayed below are the off-line of scrimmage linebackers who played at least 500 snaps.
|1||Luke Kuechly||167||18.53%||48||Miles Burris||103||12.06%|
|2||Bobby Wagner||122||17.73%||49||Brian Cushing||68||11.91%|
|3||DeAndre Levy||152||17.25%||50||Michael Wilhoite||96||11.90%|
|4||Lavonte David||133||17.23%||51||Geno Hayes||57||11.66%|
|5||Curtis Lofton||142||16.92%||52||Avery Williamson||78||11.64%|
|6||Telvin Smith||100||16.92%||53||Preston Brown||94||11.53%|
|7||Jerrell Freeman||128||16.80%||54||Mason Foster||52||10.99%|
|8||D’Qwell Jackson||157||16.67%||55||Perry Riley||79||10.96%|
|9||Rolando McClain||91||16.46%||56||Donald Butler||59||10.33%|
|10||Brandon M. Marshall||117||16.18%||57||Larry Foote||86||9.89%|
There aren’t any huge surprises at the very top, though Telvin Smith’s name does jump out at No. 6. It’ll be interesting to see if he can carve out an every-down role this season with Paul Posluszny back in the mix. I should also note that Chris Borland did post a higher tPOP than Kuechly, but he didn’t meet the qualifying snap cutoff for this list.
At the bottom end of the 57 qualifiers, we see some names who floundered for IDP purposes last season – Brian Cushing, Mason Foster, Perry Riley, and Donald Butler. Their tPOP numbers are a good indication as to why that occurred.
Avery Williamson and Preston Brown also stick out at the bottom. Both are young players who have garnered some hype in IDP circles. However, they’re going to need to improve on these numbers in order to establish themselves as viable fantasy options. This wouldn’t be an unprecedented feat. While most observed players tend to post very consistent tPOP numbers, there are some like DeAndre Levy, who improved significantly from one season to the next. Levy’s tPOP was a mere 13.66 percent in 2013.
Predicting those jumps in productivity will be extremely challenging, but tPOP gives us a strong baseline in terms of what we can expect out of our IDPs. At this point, we’ve only scratched the surface of what we can do with tPOP. In the coming months, I plan to further flesh out this metric.
Jeff Ratcliffe is the Assistant Managing Editor and resident IDP maven and DFS junkie of PFF Fantasy.