After a weekend of Twitter Percy Harvin Fantasy Football Hot Takes, I did my own investigating into the Percy Harvin trade. Here are my insights as they relate to his impact on the Jets. I say “insights” and not “conclusions” because frankly there aren’t any conclusions to be made yet. Let’s get that out of the way first. If you’re looking for “Harvin + Geno Smith = X Fantasy Points/Game”, that’s not what I have to offer. For one thing, this will be a very fluid situation. Learning a new offense mid-season is a big challenge for Harvin; likewise, learning Harvin will be a big task for Geno Smith and the rest of the offense. Presumably their performance improves as the season progresses, but I don’t think we can make any definitive statements yet.
But I do think we can re-visit the type of player Harvin is, look for general game situations he might impact, and take a look at how he’s affected the performance of other players in the past. At the very least, you’ll be able to confront the next Hot Take with some actual information. And hopefully you’ll have a good foundation from which to start formulating your own conclusions.
This is a good move for the Jets. Percy Harvin is a good real life football player. RotoViz has generally not been a fan, because he’s an inefficient producer of fantasy points. But as far as the actual game goes, he’s good. How good is he? Depends how you look at it, but here are three different views.
1. According to Advanced Football Analytics, for all receivers from 2009-141 Harvin ranks 13th in total expected points added, and 17th in win probability added. That’s very good. Take away his name, and we’d be excited to add that type of player to our team.
2. Using our own Fantasy Efficiency App, I culled a set of comparable players. I focused on players with similar receiving market shares and career targets.
That’s good company. Not “Hall of Fame” company, but good nonetheless. These numbers are also just receiving based, and don’t include his rushing production. Again, take away the name “Percy Harvin” and swap in any of those other names2 and we’d be pleased with the addition.
3. Using Pro Football Reference’s Approximate Value measure, Harvin has a per-game rate of 0.7. That’s a very good number. It’s also better than any current Jets skill position player. The three most targeted Jets so far this season3 are Eric Decker (o.48), Jeremy Kerley (o.33), and Jeff Cumberland (o.18). Their average Approximate Value/Game is 0.33. Swapping Harvin for Cumberland brings the average to 0.5. That’s a big boost.4 Apart from any specific stats he generates, Harvin’s presence on the field raises the quality of the Jets’ offense substantially.
Nuts & Bolts
So Harvin is an obvious talent upgrade to the Jets’ roster. What might that look like on the field? There are three specific areas where Harvin’s skill set could be beneficial.
1. The first is on long yardage situations. This table compares Harvin’s career performance to the 2014 Jets in situations where there are 10 or more yards to go.
Harvin’s 317 plays include 249 targets and 68 rushing attempts
The Jets are currently dead last in the NFL in first down percentage and yards/play on plays with 10+ yards to go. Even if Harvin converts long yardage situations into first downs at half his career rate, it will be a big help to the Jets’ offense. That’s a lot more sustained drives.
2. The second area is in the red zone. Maybe. Take a look at this table, which shows the red zone TD rates for the Jets’ skill players.
|Rec RZ TDR||16.1%||0.0%||11.1%||27.8%||21.4%||40.3%||30.0%||25.0%|
|Rush RZ TDR||20.8%||13.4%||16.0%||n/a||0.00||n/a||n/a||n/a|
Harvin’s not actually that good as a receiver. Ex-Jet David Nelson handily tops him, as does the diminuitive Jeremy Kerley. But as a rusher, Harvin’s red zone TD rate is excellent. If the Jets are willing to use Harvin in this manner, he could provide a boost to their offense.
3. The third benefit Harvin could bring involves boosting Geno Smith’s performance. This table presents data comparing quarterback performance with and without Harvin.
|Favre – Harvin||171||64.9||8.3||6.0%||97.5|
|Ponder – Harvin||148||74.3||8.4||5.4%||103|
|McNabb – Harvin||32||71.9||6.9||0||90.6|
|Wilson – Harvin||27||85.2||5.6||0||74.4|
In the case of Donovan McNabb and Russell Wilson, the sample size of passes thrown to Harvin is very small, and in the case of Brett Favre, the number of targets to Harvin vs. the total number of passes in his career is very small. The most illustrative example here is probably Christian Ponder. His career most closely overlaps Harvin’s. Also, Ponder, like Geno Smith, was a young, maybe not very good QB. Harvin was a boon to Ponder’s performance.
In every case, QBs have had a better completion percentage and lower interception rates throwing to Harvin. I think that’s important. Incomplete passes don’t do anything to help an offense, so the more completions the better. Favre and Ponder also posted better yards/attempt and TD rates, and all except Wilson posted better QB ratings when targeting Harvin.
So, more first downs, a potential boost in the red zone, and maybe some improved quarterback play. Sounds good. It’s hard to be more specific…but I’ll try.
This is an intriguing situation to me. Superficially, it seems that a player like Harvin who can be dangerous near the line of scrimmage would help Decker get more favorable situations down field. There’s also a somewhat useful historical comparison to be made: Sidney Rice. From Mock Draftable:
In addition to being physically similar, both also sport similar yards/target and yards/reception numbers. Finally, Rice has a reasonably sized set of games with and without Harvin to analyze.
Those numbers come from 2009 to 2011, when both Rice and Harvin played in Minnesota. Those are some nice bumps, especially in the yards/game department. Superficially this fits with the idea that Harvin can help a bigger/downfield receiver.
Chris Ivory/Chris Johnson
The news here might not be as good. Consider the splits for Adrian Peterson and Marshawn Lynch with and without Harvin. Lynch’s numbers are from 2013 and 2014; Peterson’s from 2009-12, when he played with Harvin.
Both backs performed better as receivers when Harvin was in the game. But both saw fewer rushing attempts and yards. The Jets backfield workload has been trending in Ivory’s direction this season, and I think that Harvin is more a threat to Johnson’s workload. They’re similar physically5 for starters. But that’s just a guess. It’s also possible that putting Harvin and Johnson on the field together could allow Johnson some big play opportunities that offset a reduction in total workload. Neither of Harvin’s former teams employed a backfield committee so it’s hard to say what might happen here. My best thought is that Ivory can do things that Harvin can’t do, while Harvin can pretty much do anything Johnson does. So I’d expect Ivory’s workload and production to stay roughly the same, while Johnson’s workload declines quite a bit.
Jeff Cumberland/Jace Amaro
There’s not a lot of games here to look at, but here are Kyle Rudolph’s splits with and without Harvin, while both were in Minnesota.
On a smaller sample of 16 total games (4 with, 12 without), Zach Miller also posted lower numbers across the board with Harvin in the game. That wouldn’t be a bad thing for the Jets. Among tight ends with 10 or more targets, Amaro ranks 21st in per-target efficiency. Cumberland is posting a brutal, 49th best -0.51 per target score. As with the running backs, I’m just speculating, but it seems to me that Harvin will take work away from Cumberland, while Amaro’s workload stays relatively flat. That would boost the efficiency of the offense as a whole.
Again, the conclusion is that we don’t really know what will happen. A lot depends on how quickly the Jets can integrate Harvin into the offense, and how much of Harvin’s skill set and desire remain intact. But in my personal leagues, I’m starting with the expectation that Jeff Cumberland, Jeremy Kerley, and Chris Johnson are losing work, Chris Ivory is relatively unaffected, and Eric Decker and Geno Smith are good bets to get more productive. It also seems reasonable to assume that the Jets will make every effort to get Harvin heavily involved. But the Jets’ remaining schedule is relatively neutral for WRs over the rest of the season, and downright difficult in the fantasy playoffs (sixth most difficult). I’m not sure we can count on Harvin for a big impact rest of season.
- Excluding 2013, when he was injured. (back)
- Okay, maybe not Bill Schroeder. (back)
- Excluding Jace Amaro, who doesn’t yet have an Approximate Value. (back)
- For some background on how to think about AV/G and receiver valuation, see this article. (back)
- Harvin is 71″, 192 lbs, 4.39 40, 158 Explosion; CJ2K is 71″, 197 lbs, 4.24 40, 165 Explosion. (back)