Fantasy: House Rules - The Quest for an IDP Scoring System, The Findings
Collaborative effort by Ross Miles & Jeff Ratcliffe
After nearly two months of data analysis and in-depth email discussions, we have arrived at our PFF Fantasy IDP house rules. It was by no means an easy process, but the goal was, and still is, rather simple – To create a standardized IDP scoring system that is both reflective of actual game play and does not either under- or overvalue IDPs compared to their offensive counterparts.
To begin, the IDP crew, which consists of Ross Miles, Kevin Greenstein, and yours truly, collaborated through email regarding which statistics we felt should be included in an ideal IDP scoring system. Fairly quickly and unanimously, we agreed that we must include – solo tackles, assisted tackles, sacks, safeties, blocked kicks, tackles for a loss, interceptions, passes defensed (or deflected), forced fumbles, and fumble recoveries. Most of these statistics are standard in all IDP scoring systems, but we also decided to use the less common tackles for a loss and passes defensed. We collaboratively felt that the inclusion of these stats would help to increase the fantasy value of the defensive backs and defensive linemen, who are often undervalued by many IDP scoring systems.
With our statistics in place, Ross and I then set out to do some analysis. In my work, I took a look at the frequency at which each key statistic occurs. Here, I established my “solo ratio,” which is a metric that shows how often each statistic is recorded compared to solo tackles. Ross then set out to benchmark IDP data in order to gauge IDP production. In the first part, Ross uses a basic IDP scoring system that you may find many leagues using, and applies that system to all of the IDP data here at PFF.com from 2008-2010. His results are eye-opening, as LBs dominate fantasy scoring in that system. In the second part, he elaborates on the inherent problems of using a basic system, as ILB3s actually outscore DE1s by an average of 9.5 fantasy points per season. If we apply that to last season, that would mean someone like Karlos Dansby outscoring Justin Tuck!
So with our scoring system, we want to set right these fantasy wrongs. To do so, Ross and I independently developed our own IDP scoring systems. In this two heads are better than one approach, we wanted to see what each other would come up with and then debate the merits of each approach. We knew we would certainly have overlap in some areas, but also points of departure in others. In the places where we disagreed, we then discussed the logic and data behind each other’s decisions. This collaborative approach led to what we feel is a very effective and statistically grounded system that accomplishes our initial goals. In this article, we reveal to you our thought process for each statistic and our collaborative conclusions. Before we get to that, we felt it important to first discuss our motivations and personal philosophies regarding IDPs in fantasy football.
Jeff – My motivation for developing a standardized scoring system stemmed from a frustration over my many years of playing in IDP formats. Invariably, LBs were the highest scoring players in every league. Now, I understand that the performance of someone like Patrick Willis in 2009 should rightfully lead IDP scoring, but I was in one league where LBs accounted for each of the top-12 scorers and 15 of the top-20. Charles Woodson, the Defensive Player of the Year, ranked 14th in fantasy scoring in that league, and was outscored by David Hawthorne, Barrett Ruud, and James Laurinaitus. Not to knock any of those three guys, but something just isn’t right here.
This league used what I’m going to call the “typical IDP scoring system.” In this system, tackles received one point, assisted tackles a half point, sacks and interceptions got two points, and forced fumbles and fumble recoveries were one point each. This system also scored touchdowns at six points and the rare blocked kicks and safeties at two points. The typical scoring system allows for what are essentially LB2s to outscore the league’s best safeties, and top defensive lineman don’t even sniff the top-50 overall. But is this reflective of actual play? Are LBs more valuable to their team than any other defensive position? Anyone who knows football will tell you that they’re not.
So my goal is simple, to develop an IDP scoring system that is more reflective of actual on-the-field play. I understand that due to statistical limitations, it is impossible to completely mirror real football action in fantasy football. Shut down corners have their best days when nothing happens, and this does not translate well to the world of fantasy football. But I still feel we should strive to get as close to reality as possible. As I have already pointed out, most IDP scoring systems tend to overvalue the LB position while devaluing the D-Linemen and D-Backs. Ultimately, I want to move away from the typical IDP scoring system towards a system that is more reflective of what happens on the field and that promotes more parody between the three defensive levels.
Ross – One of my primary motivations in creating a revised IDP scoring system was to reward players for game-changing performances fairly. When I see Justin Tuck having games with multiple sacks and turnovers I wanted his fantasy score to reflect that. My feeling, like Jeff, was that traditional scoring systems often overvalued LBs, ILBs in particular, to the extent that a typical IDP strategy was to stockpile them on your roster over other positions.
Not only is my goal to create a more representative scoring system, but one that works in harmony with typical fantasy scoring systems for offense too. A system throwing out IDP players that can compete with QB1/RB1/WR1s would be excessive in my opinion, because it is the offense in the NFL that scores the points and wins the games, so it should be in fantasy football too, but a system with closer parity needs to exist.
Another goal is to find a way to make DBs, especially CBs, fantasy relevant. One issue is that the better a shutdown corner is, the less opportunities they have to make a play (and achieve a stat). The prime example is Nnamdi Asomugha, who has been targeted an average of 29 times a year since 2008. However, Asomugha seems to be an outlier, as over the over the same period Darrelle Revis has been thrown at 92.7 times a season, and Charles Woodson 82.7, so there should be a way that we can improve the fantasy output of elite corners.
Breakdown by statistic:
Ross – (0.75 points) Solo tackles are the basis of any scoring system and are the main reason LBs are the most valuable IDP players. Standard systems value tackles as 1 point, but I’d prefer to lessen that and see more emphasis placed on turnovers to reward impact players.
Jeff – (1 point) I also view solo tackles as the baseline for an effective IDP scoring system. They are far and away the most commonly occurring defensive statistic. In fact, a solo tackle was recorded on 79.4% of all snaps over the last three season. I prefer to keep tackles at the standard 1 point, as some host sites are not as customizable as others. This way, the scoring system can be applied to all host sites, even those who do not allow for fractional points (other than .5 points).
Collaborative – (1 point) In an effort to make our IDP scoring system as standardized and accessible as possible we should stick with solo tackles being worth 1 point.
Ross – (0.5 points) Assisted tackles are a sign that a player is in and around the action. If I have a player making high numbers of them, I want him to be re-numerated for playing an active role in shutting down the offense, but not over-compensated.
Jeff – (0.5 points) I agree with Ross here. An assisted tackle was recorded on 21.8% of defensive snaps from 2008-2010. Though they were less frequent than solo tackles, I feel their value should be roughly half that of solo tackles due to the fact that the IDP is not making an individual effort in recording an assisted tackle.
Collaborative – (0.5 point) We are both in agreement that assisted tackles should be worth 0.5 points.
Ross – (3 points) Sacks are the most common form of turnover, more than twice as likely to occur than a interception or forced fumble. They also count as a tackle, and are given an ‘extra point’, so need a slightly reduced value in comparison to other turnovers to prevent over-valuing pass-rushers.
Another factor in my valuation of a sack is the fact that a good pass-rusher will reach double figure sack totals for a season, then will supplement his scoring with tackles for loss, which as we discuss below, should also be worth 3 points. Over the past three years 48 players have 15 or more combined sacks and TFL, while 187 have had 10+ in a single season, so I’m wary of putting too much emphasis on them. Placing too high a value on a sack can also also give one dimensional players and situational pass-rushers a boost that increases their value.
Jeff – (4 points) As Ross points out, sacks occurred on 3.2% of all snaps over the last season, while interceptions and forced fumbles both occurred on just 1.5% of snaps. That means a top-tier DE may record just one sack per game. Scoring sacks at two points, as most typical IDP systems do, severely devalues DEs. To restore parody between defensive levels, I suggest 4 points for a sack. Now, remember that a sack is also recorded as a tackle, so this is actually a 5-point play. While that may seem like a lot, remember that most DEs do not even top 40 solo tackles in a season, and only the elite with top 50 solo tackles.
The importance of this shift can be seen, for example, with Jared Allen’s 2009 performance. Despite finishing the season with the second-highest sack total of 15, Allen ranked 59th in fantasy scoring among IDPs in typical scoring systems that scored sacks as two points. If we bump sacks to three points, Allen jumps to 20th in fantasy scoring. At four points, Allen comes in at 6th in IDP scoring. I feel this is much more reflective of his monster year than 59th.
Adjusting sacks to 4 points will also also serve to increase the fantasy value of outside linebackers in a 3-4 defensive scheme. Let’s take a brief look at Clay Matthews from last season. Nobody can argue that he did not have impressive season, but in typical IDP scoring systems he ranked 58th overall despite recording the fourth highest sack total in the league. In fact, all of the top-4 in sack were 3-4 OLBs – DeMarcus Ware, Tamba Hali, Cameron Wake, and Mattews.
The problem with 3-4 OLBs from a fantasy perspective is that they play statistically much like 4-3 DEs and typically do not record a lot of tackles. Of the four mentioned 3-4 OLBs, Ware led with 57 solo tackles last season. By adjusting the value of sacks, Matthews jumps from 58th to 31st in fantasy scoring, Ware goes from 38th to 14th, and perhaps most impressively, Wake jumps 83rd all the way up to 28th. I feel these adjusted rankings are much more reflective of actual play.
Collaborative – (4 points) Our system was inspired by two main goals, to reflect on field performance and to reward playmakers. By taking Jeff’s value for sacks it makes sure players like Justin Tuck or DeMarcus Ware get the points, as well as plaudits, they deserve.
Tackles For Loss
Ross – (3 points) A tackle for loss is a big play for a defensive player and should be rewarded as such. The only difference to a sack is the fact the player being tackled isn’t the QB. It always frustrates me when an IDP system ignores this stat as 3-4 DEs and penetrating DTs will often make their mark by making tackles for loss. A successful play for them often does not result in a sack, as their role in the gameplan is to prevent the opposing team having a successful running game. And what about those great tackles by a blitzing nickel corner, who changes his read and makes a great play on a RB in the backfield? 3rd and inches and your starting LB stuffs a fullback run up the gut and you get only a point? I want to see a system that rewards players for contributing to a successful defensive unit, executing their role superbly, whether it is a highlight reel sack or not.
Jeff – (3 points) Ross makes an excellent point here, and I agree with him 100 percent. However, I value a tackle for a loss at less than sack. This is primarily because tackles for a loss occurred more frequently at 4.3% of all snaps over the last three season. As such, they should be valued slightly lower than sacks.
Tackles for a loss are perhaps the least commonly used IDP statistic, but these have tremendous impact on the game. Last season, Chad Greenway led all players with 14 tackles for a loss. In typical IDP scoring systems, which do not include tackles for a loss, Greenway finished the season 32 fantasy points behind leader Jerod Mayo. While their solo tackles total was relatively similar at 114 for Mayo and 109 for Greenway, Mayo recorded 26 more assisted tackles, one more sack, and two more fumble recoveries.
While that certainly sounds like Mayo deserves to be the higher fantasy scorer, consider that Greenway recorded eleven more tackles for a loss than Mayo. In fact, only eight players recorded double-digit tackles for a loss last season. This stat deserves to be put into its proper fantasy context. So if we add 3 points per tackle for a loss to the typical system, Greenway actually outscores Mayo by eleven points. While this is eyeopening, keep in mind that we will also be adjusting the typical system in terms of sacks, forced fumbles, and fumble recoveries, so that Mayo would again come out ahead of Greenway in scoring. However, the gap would be much closer than the typical IDP scoring system, and thus more reflective of on the field play.
Collaborative – (3 points) Another stat we agreed upon in value and in reasoning.
Ross – (8 points) A rare play that has real significance on a game. Scoring points on defense and getting the ball back for your team can be a huge momentum shift in a real game.
Jeff – (10 points) While I know it’s crazy to value any statistic in the double digits, consider not only the impact on the game that Ross mentions, but also that safeties occurred on just .05% of all snaps over the last three seasons. That’s one safety every 2128 snaps. In a typical season, a team will play roughly a little over 1000 snaps, so the odds say that a team can go through two whole seasons without recording a single safety.
Collaborative – (10 points) We both agree that safeties are extremely infrequent and have a big impact on a game, so we have elected the higher value. Playmakers will be rewarded.
Ross – (5 points) Whereas the elite DEs can easily have double-digit sack seasons (45 players since 2008), no player has had more than 9 interceptions. On that basis interceptions need to be worth more. I don’t want to over-value them though with my system looking for ways to reward DBs as a group, not just the ballhawks.
In my system Charles Woodson phenomenal DPOY campaign in ‘09 would have been good enough to be the #1 IDP scorer that year, and the 2nd best season since 2008. This increased value would have also helped Darren Sharper rank as the 11th best fantasy player in 2009 with his 9 interceptions.
Jeff – (6 points) Ross hits the nail on the head. Interceptions are less than half as likely to occur as sacks. However, I too do not want to overvalue ballhawk CBs like Asante Samuel. Instead of doubling my sack value, as the numbers say I should, I chose six points.
Like sacks, this shift has significant impact on IDP player value, particularly with the DBs. As I mentioned earlier, Charles Woodson’s stellar 2009 season ranked out at 14th among IDPs in typical scoring systems that score interceptions at two points. If we bump interceptions to six points, Woodson finished second in fantasy scoring among IDPs behind only Patrick Willis, who had a stellar 2009 campaign of his own.
Collaborative – (6 points) As we rounded out our valuations that meant pass deflections would be worth only 1 point, so to counter the impact of that on defensive backs we decided to value interceptions at 6 points.
Pass Deflection (or Passes Defensed)
Ross – (1.25 points) By giving a slightly higher than standard value to passes defensed it helps to increase the value of CBs who often struggle to accumulate gawdy stats in other areas. I also think a pass defensed is worth more than a tackle as it prevents the offense from gaining positive yardage, and can also be touchdown saving if in the endzone, where a tackle makes no difference.
Creating a system that can reward elite play from DBs is difficult, because the better they are at their job, the less targets and opportunities to makes interceptions, passes defensed or tackles. Last season saw the emergence of Brent Grimes into the public eye for the first time with some excellent performances. Grimes had 23 pass deflections and made the Pro Bowl, yet in regular scoring system without pass deflections he was only the 79th best player.
Jeff – (1 point) I understand Ross’s point, and I can’t say I disagree. However, I again want to avoid a scoring system that will work on some host sites, but not in others. Pass deflections are the third most frequently occurring defensive fantasy statistic behind tackles and assisted tackles. At an 8.3% clip over the last three seasons, pass deflections were roughly 2.5 times more frequent than sacks.
Unfortunately, many leagues do not use pass deflections in their IDP scoring systems. This has a tremendous impact on the defensive backs, especially the corners. Take for example, the case of Darrelle Revis. Despite posting a ridiculous 35 pass deflections in 2009, the typical IDP scoring system ranks Revis 141st overall and 21st among CBs behind the likes of Derek Cox, Sheldon Brown, and Dominque Rodgers-Cromartie to name a few. If we add one point for each pass deflection to the scoring system, Revis jumps to 63rd overall and the 5th ranked CB and we’re getting much closer to mirroring the on the field performance.
Collaborative – (1 point) As mentioned above, we want this system to be implementable on as many fantasy host sites as possible, so we went with pass deflections valued at 1 point.
Ross – (5 points) As I want my system to reward big plays I have put an emphasis on forced fumbles. They can kill a drive and change the course of a game. I also view it as an ability that can be worked on and improved. A case in point would would WR and special-teamer Kassim Osgood who has 12 in the last three years, one less than DeMarcus Ware and one more than Dwight Freeney.
Jeff – (3 points) Like Ross, I feel that forced fumbles tend to be undervalued in typical IDP scoring systems. These are often game-changing, and occur on just 1.5% of snaps. To score them the same as solo tackles sells short the fantasy value of players who force a lot of fumbles. As Ross points out, this is a skill, and certainly players are darn good at forcing fumbles. Case in point – James Harrison.
Over the last three seasons, Harrison has forced 18 fumbles. He’s a master of the strip-sack. In 2008, he recorded seven forced fumbles. In typical IDP scoring systems, Harrison ranked 5th overall in fantasy scoring. This system puts him ten points behind leader Patrick Willis. However, if we change forced fumbles from one to three points, Harrison actually outscores Willis by 8 points because Willis recorded just one forced fumble. So while the typical system rewarded Willis for accumulating more solo tackles than Harrison, the adjustment to forced fumbles underscores their in-game importance.
Collaborative – (4 points) A happy medium between our independent scoring systems. It makes sure that the game changing nature of the play is reflected in a player’s fantasy score.
Ross – (2 points) Although recovering a fumble would appear to be a big play, it often involves an element of luck. I don’t feel right giving a player more than 2 points for picking up a bouncing ball, regardless of which team ends up with the ball and any momentum shift on the game.
Jeff – (3 points) Ross is absolutely right about the element of luck involved. Unlike forced fumbles, where we see consistence from players like Harrison, fumble recoveries are a bit of a crap shoot. In fact, no one player has recorded 3 or more fumble recoveries in two or more seasons over the last three seasons. But does that mean the stat is any less important?
Here’s where I do disagree, because a forced fumble has no value if the offensive team recovers. From 2008-2010 there were 501 forced fumbles, but only 339 fumble recoveries. That means the offense recovered the ball roughly one-third of the time. In these cases, the forced fumble was not game-changing after all. So while recovering a fumble does not depend on as much skill as forcing a fumble, it is extremely important to the flow of the game and should be valued equally.
Collaborative – (2 points) The argument that fumble recoveries carry a large element of luck wins out here.
Ross – (6 points) They are a rare occurrence, but I don’t feel they are as game changing as a safety, or as skillful (and repeatable) a play.
Jeff – (6 points) These are extremely rare, with one occurring every 711.8 snaps. They can have a significant impact on game momentum as Pat Chung showed us last season with his two blocked kicks against the Dolphins. But Chung also showed us that these are not, as Ross says, very repeatable, as his two blocked kicks against the Dolphins were his only two of the season.
Collaborative – (6 points) Again our research and analysis led us to the same conclusion.
Ross – (6 points) Score a TD on offense and you get 6 points, so why not on defense too. The play itself will also be boosted by points for a fumble recovery or interception, so that gives it the increased value it deserves.
Jeff – (6 points) Despite defensive TDs occurring much less frequently than offensive TDs, a TD is a TD regardless of when it occurs.
Collaborative – (6 points) Another agreed valuation.
The PFF Fantasy IDP Scoring System
|Solo Tackles||1 point|
|Assisted Tackles||.5 points|
|Tackles for a Loss||3 points|
|Passes Defensed||1 point|
|Forced Fumbles||4 points|
|Fumble Recoveries||2 points|
|Blocked Kicks||6 points|
So there you have it, the official PFF IDP house rules. Moving forward, our IDP rankings and projections will be based on this scoring system. Before we get to those, however, we will take this scoring system and apply it the IDP data from the last three seasons, much like Ross did in his benchmarking work. We will be sharing that data with you in the next few days, so keep your eyes peeled.