Trimming the Back End of Your Roster

Peter Howard

Editor’s Note: This article is submitted by a new Member Corner writer, Peter Howard.  We look forward to seeing much more of Peter’s work in the future.

It’s that time of year. The NFL draft is over, and trades are starting to fly (in leagues where trades fly.) It is still a good rule to allow three years to fully understand what a player can be in the NFL. But while market value may be the best arbiter on the trade market, it shouldn’t be for managing the back end of your roster.

In general, players with higher draft capital get their opportunity early. This means in most cases the odds drop on these players the further away from their draft year we get, while the numbers remain steady for later round prospects. But if I can’t convince you to cut Nelson Agholor ahead of Wendall Williams and Roger Lewis then I at least want to encourage a re-evaluation process for the bottom end of your roster.

To do that we should try to balance unpredictability by focusing on probability. In that vein, here are some tips for re-evaluating the later round wide receivers on your roster.

What have later round breakouts had in common?

Honestly, not much. If we assume that you’re more likely to think about cutting players with less draft capital it might pay to take a closer look at the ones who have the least, but still broke out.

The table below shows information on players drafted in the last four years. Point totals are for typical PPR scoring leagues.

production by round table corrected

Wide receivers drafted in the first round of the NFL draft still score more fantasy points. But the 8,000+ points from the later round players is an intriguing pool to mine and not one we should overlook.

If we just look at the unlikeliest of unlikely producers in the biggest player pool we should find some clues as to what might help. The top ten undrafted players from the last four years are; Allen HurnsWillie SneadTaylor GabrielAdam ThielenTyrell WilliamsCorey BrownMarlon BrownSeth RobertsAlbert Wilson and Cameron Meredith.

People disagree on how to value these players in 2017. But if you want to mine fantasy points from undrafted prospects, these were the players to hit on since 2013. Even within this group, the most valuable players – by ADP – share certain common traits.

Routes to success

What can we see in this group? Is there any commonality to the routes they took to find success in the NFL?

Players on the same roster they originally signed to include; Hurns, Snead, Thielen, Williams, Roberts, Wilson and Meredith.

Players who have changed roster include; Gabriel, Corey Brown and Marlon Brown.

You may not like Allen Hurn’s prospects in 2017. You may value Taylor Gabriel more. But if a team keeps a player with poor (or no) draft capital, it increases the probability they breakout out in a meaningful way. This makes sense when you think about it – why keep a player you don’t have anything invested in? They must have proven themselves in order to stick around.

I don’t know who is on the back end of your roster. But I suggest dividing them into a list of players who have stayed on the same team since they came out of college, and those who have not. It won’t be foolproof, but as a deciding factor, it sides with probability over less trustworthy narratives.

Athleticism

Something you should not use to trim your roster is athleticism. It’s not possible to produce a guideline by athletic measurements from undrafted breakouts. The difference between Tyrell Williams and Willie Snead alone will inflate the range to such an extent as to make it useless. Our friends over at Player Profiler help us illustrate this:

screen shot 2017 05 09 at 09.08.37

screen shot 2017 05 09 at 09.09.09

So cutting your least athletic players with low draft capital isn’t the answer. But that doesn’t mean metrics abandon us when mining less likely producers.

Metrics

Numbers are still on our side here, but the range is broader and a ten player sample is small. This means we have to be more tentative with our predictions and ranges. However, there are some I think we can pay attention to safely. Looking back at Snead and Williams’ profiles provides confidence that Dominator Rating and Breakout Age are still useful indicators.

If we eliminated Marlon Brown, Gabriel and Corey Brown – who we have already highlighted as bigger outliers – then we find that the lowest College Dominator rating in our list is Hurns at 28.8% (46th percentile) and the oldest Breakout Age is Seth Roberts’ 22.5 years old (seventh percentile). The median College Dominator is 37.35% (above the 68th percentile) and the median Breakout Age is 20.7 (above the 47th percentile.)

So while the range is broad, College Dominator becomes a better indicator then Breakout Age. College Dominator remains significantly above average while Breakout Age (primarily thanks to Meredith and Williams) is significantly below the average.

These are two predictive metrics for success in the NFL. They may give you a place to start cutting your rosters, no matter what a player’s draft capital looks like.

Market Share

The work done by Jon Moore and Jim Kloet at Rotoviz on age-adjusted production is not just great, it’s useful for backdating our opinions. If you have not considered age-adjusted Market Share to evaluate rookie prospects in 2017 I would encourage you to start. Looking at our rosters we can also project some of this process back to prospects drafted in the last four years.

For my part, I found myself with a lot of Kenny Bell and Rashad Greene (don’t laugh) in deeper leagues. I liked Bell’s athletic profile and I liked Greene’s counting stats (two 1,000+ yard seasons in his final two years at college.) So I did my own, inferior, version of Jim Kloet’s visualization to re-evaluate.

The age in these graphs indicates the age players were at the start of each season. The benchmarks for second and third round “hits” are based on the tireless work of our rotoviz friends and indicated by the yellow and brown horizontal lines.

graph rashad greene production

graph kenny bell production

Both were fairly productive young, which is nice. However, 21.8 and 17.8 MSYDS aren’t overly impressive even for younger players. Neither prospect met the average of successful second and third round draft prospects either. These two players went in the fifth round of their drafts, and it made the decision to cut them over others much easier for me.

It can be tedious work, tracking down rookie stats from previous years, and so can making graphs. If there’s anyone in particular you’d like to see a graph on you can contract me on Twitter. I can’t guarantee I have it or how long it will take me to get to it, but I have developed an inexplicable fascination with making them and would be happy to try and help.

To be clear this isn’t the ultimate indication a prospect is doomed to fail. But if I’m making cuts for roster space I’d rather use new processes in understanding past prospects than deciding whether or not I “feel” like it’s ever going to happen.

How long should I wait?

The longest hibernation period in our sample so far (from fantasy irrelevance to fantasy starter) is Adam Thielen. He was in the NFL for three years before he was worth starting (or rostering.) Allen Hurns had some immediate impact in his rookie year, before – like a groundhog seeing his shadow – scampering back onto fantasy benches. Cameron Meredith and Tyrell Williams did nothing in their first year but slowly found opportunity in their second.

In short, there’s no one timeframe. But if you have rostered a player with an interesting profile for a year (looking at you Chester Rogers, Wendall Williams, and Roger Lewis, owners) then you shouldn’t have expected anything. And if you own Robby Anderson, you should be pleasantly surprised, and encouraged, he did something in 2016.

More importantly, a player with no draft capital who meets our earlier requirements has already eaten into his hibernation period. In this way, we should be more interested. A player a year into the league has already put at least some of their potential hibernation period behind them.

Conclusion

There are warning signs for lower round prospects we thought might last. Likewise, there are some positive trends that should embolden us to keep holding. Not all of them are intuitive. So to bet on the probabilities follow these guidelines:

  1. Don’t give up on a player just because they don’t have uber athleticism
  2. Be more willing to drop players who have switched teams
  3. Use Market Share and new scouting techniques to re-evaluate older sleepers
  4. Don’t cut a player just because he’s been on your roster longer
  5. Don’t keep a player because he entered the league more recently
  6. If you want to see a graph, head to Twitter! (@pahowdy)

In some ways, last year’s sleepers should be more attractive than prospects from this year. Opportunity comes later for lower drafted prospects. First round picks become less likely to breakout if they don’t show anything encouraging in their first seasons. However later round prospect’s probabilities hold steady and their outlook can improve even before they break out. That’s if they are going to breakout at all…

peter howard
Follow me