clock menu more-arrow no yes mobile

Filed under:

A Probabilistic 2020 Redskins Roster Projection, Part 1 of 2

Predicting the Unpredictable

NFL: DEC 15 Eagles at Redskins Photo by Mark Goldman/Icon Sportswire via Getty Images

In the lead-up to this year’s draft, I published an article on Hogs Haven highlighting the unpredictability of draft outcomes. I described the draft as a probabilistic system, meaning that NFL career outcomes of draft prospects are only partially predictable. The two main reasons for that are: (1) it is not possible to precisely nail down all of the known factors that contribute to NFL success ahead of the draft, and (2) random events and factors outside of anyone’s control, such as injuries, play a major role. The same is true for predicting an NFL season.

We are heading into prime roster projection season on Hogs Haven. In the coming weeks, the many highly knowledgeable expert contributors will pour through tape breakdowns, draft profiles, media analysis, and sometimes even reports from sources within the Redskins organization to put together projections for what the opening day roster will look like.

In the seven or eight seasons I have been avidly following the site, I’ve noticed that these preseason projections often tend to describe a better team than the one that takes the field on opening day. Why is that? Two reasons come to mind. For one, they are often built on optimistic assumptions. All of the early to mid-round draft picks live up to the upside of their scouting profiles. All of the injured players come back to pre-injury form. Every player with off the field issues cleans up his act. Basically, every question mark resolves as a positive. And second, players never get injured.

On the other side of the coin, the normal process of making a roster projection also misses out on the real excitement of the NFL preseason. There are no surprises. That’s not a criticism. No serious analyst, putting together the results of their hours of film review and other research, would think to project an undrafted free agent WR converted to a CB in camp to make the roster, let alone start two games his rookie year. And who, other than Mark Tyler, would have predicted that Giants draft bust OT Erek Flowers would not only start at LG, but become arguably the best player on an offensive line that includes Brandon Scherff? These types of surprises happen nearly every year, and are a big part of what makes the NFL preseason football so much fun to follow.

This season might well be even more unpredictable than ever due to the disruption caused by the coronavirus. With the elimination of OTAs, minicamps and possible truncation of training camps and preseason, teams will likely have less ability to evaluate rookies, returning players and incoming free agents before the season starts, resulting in more guesswork than usual.

In this two-part miniseries, I will use historical NFL datasets in an attempt to generate a 2020 Redskins roster projection that adds an element of realism that is usually missing. It will contain both pleasant and unpleasant surprises. To be absolutely clear, I am not attempting to improve on the accuracy of roster projections. Rather, I am attempting to improve on their realism by introducing the missing element of unpredictability. This is not an attempt to predict the final starting roster. Rather, I am aiming to generate a plausible roster scenario, which mostly aligns with general expectations, but contains a few surprises that no one would have predicted.

In the first article of the series, I will simulate a competition for open starting spots and random occurrence of preseason injuries to drive construction of an opening day 53-man roster. In the second article, I will let my Sims (pun intended) play and examine how the roster and win-loss record evolves as the season progresses.


NFL Combine - Day 2

Roster Projection Method Overview

In a stark reversal from the previous regime, the front office and coaching staff are now providing very clear and consistent messaging. Two themes which have been emphasized over and over are competition and a focus on evaluating and developing young talent. In keeping with these themes, my approach to resolving uncertainties on the starting roster will be driven by a simulated competition for available starting spots amongst the young talent on the roster, specifically players in their first three years in the league.

My approach is primarily driven by selection of the starting roster for reasons that I’ll explain in a minute. I have adopted a hybrid approach to generate the roster projection. I will start in the usual manner, by pre-populating the starting roster with the known quantities, established players and free agent (FA) acquisitions who are either “locks” or have an “inside edge” to claim/retain a starting position. I’ve stretched the definition of starting positions a little to include 3rd down back/weapon, slot receiver, slot/nickle CB, and kick/punt returner. After creating the initial roster, I will then introduce two elements of unpredictability by using probability values derived from actual NFL datasets to: (1) simulate the competition amongst the younger players for open starting positions and (2) add injuries.

Once I’ve identified the new starters, I will revert to the usual approach, by playing virtual GM to resolve any remaining uncertainties and conflicts in the starting roster. Finally, I will use my discretion as GM to fill in the backup positions with the remaining rookies, returning players and free agents.

It might occur to some readers that the idea of rookies and developmental players competing for starting spots is not really how things work. Usually, such players compete to earn backup roster spots, then work their way onto the starting roster.

My chosen approach is dictated by the datasets that are available to me. To simulate the roster competition, I used data on probabilities of drafted players earning starting spots I painstakingly compiled for a previous Hogs Haven article. Given how much work it took to pull that dataset together, I have no intention or interest in duplicating that effort to calculate probabilities of drafted players earning non-starting positions. Attempting the same thing for veteran backups and free agents would be a far more complex and labor-intensive exercise, so it’s out of the question.

From a mathematical perspective, though, the starting probability data I am using captures a whole range of events that lead to drafted players earning starting time. The veterans that got beat out when draftees earn starting spots are in there. They just aren’t the focus. And, while there may be an element of looking for your lost car keys under the street lamp to my approach, the focus on development of recently drafted players is consistent with Ron Rivera’s comments when asked recently about the possibility of signing Antonio Brown:

“If you bring in a veteran player, you’re going to stunt the growth of someone young.” Ron Rivera says when asked about Antonio Brown possibly being brought to DC. Coaches want to see what they have at each position before bringing in some vets…

-Rhiannon Walker on Twitter


Washington Redskins v Green Bay Packers Photo by Dylan Buell/Getty Images

Starting Roster Locks

Barring any huge surprises or preseason injuries, the following players are locks to make the opening day starting roster. Matt Ioannidis and Jon Allen are both roster locks, but it’s anyone’s guess which one to call the starter. I’ve punted on that one and listed them as one or the other:

C/G Chase Roullier

RG Brandon Scherff

WR Terry McLaurin

NT Daron Payne

DT Matt Ioannidis or Jon Allen

DE Montez Sweat

CB Kendall Fuller

S Landon Collins

P Tress Way

K Dustin Hopkins

LS Nick Sundberg

One notable omission from this list is Chase Young. While it would be a huge surprise to everyone if Young did not start this year, historically second overall draft picks only have a 0.78 probability of starting as rookies, leaving a real 0.22 chance that he doesn’t start. Chase will have to earn his starting position like everyone else, but you have to like his chances.

Inside Edge Players

The following players enter the preseason as the players to beat at their starting positions:

QB Dwayne Haskins

DE Ryan Kerrigan

RT Morgan Moses

Slot WR Steven Sims

RB Adrian Peterson

3rd down back/weapon J.D McKissic

CB Ronald Darby, Fabian Moreau

LB Cole Holcomb, Jon Bostic

S Sean Davis

These are either returning veteran starters, second-year players who showed starting potential as rookies, or free agent acquisitions signed to starting-level contracts. All of these players are on the inside track to retain or earn a starting position. But all can be beaten by the young players competing for starting spots. Based on comments from Ron Rivera about competition at the QB position, this includes Dwayne Haskins. However, Kyle Allen’s chances of unseating him, as a third-year UDFA, are very low.

The secondary players on this list could easily be moved into the lock category due to the lack of competition from players below them on the depth chart. But none of these players are so good or well established that they couldn’t be beat by a developmental player, such as Jimmy Moreland or Greg Stroman, staging a breakout performance in the preseason.

Some potentially controversial omissions from this list are Wes Martin, Kelvin Harmon, and Jeremy Sprinkle, who all earned starting time in 2019. Harmon does not make the cut because I don’t think he did enough to cement his starting status last year, and the team drafted a rookie WR with a similar skill set in the fourth round. Martin and Sprinkle fall into the next category.

Vacant/Open Competition

At least four positions on the starting roster are wide open at this stage. LT has no clear-cut choice to start day one. While there are returning players and FAs with starting experience at LG and TE, Ron Rivera recently declared an open competition for these starting positions, as well. As discussed above, Kelvin Harmon and fourth-round draft pick Antonio Gandy-Golden will have to compete with a host of developmental players and one new FA, pending the outcome of legal proceedings, for the second starting WR position.

In my opinion, primary kick and punt return duty (or duties) is up for grabs every year. As often happens, in 2019 Steven Sims established himself as the first dynamic return option the team has seen in many years. But he may become too valuable at slot receiver to risk on special teams.

The competition at LB is the hardest to call, due to the uncertainty regarding who might play what roles in the new 4-3 defense. There are several viable options competing for the starting spots at SLB and ILB. I have given the inside edge to returning vets Cole Holcomb and Jon Bostic for these positions. That leaves WLB. Aside from Rivera’s imported locker room leader, 38-year-old Thomas Davis Sr., who can play all three LB positions, the only starting quality player with the skill set to play WLB on the roster is Reuben Foster. If he can fully recover from his injury and clean up his act off the field, he will be a lock for starting WLB. But that seems so unlikely that I will treat him as a special case, below. Because of the uncertainty regarding the LB corps, I am going to bend the rules slightly and declare an open competition for one starting LB spot.

The Vacant/Open Competition category assumes special importance in my simulation. All young players competing for starting spots at LT, LG, TE, WR and LB get their probabilities of earning starting positions doubled. I have not included KR/PR in that bonus, because it is usually a secondary consideration for earning starting positions and it is too difficult to figure out which players qualify.


New York Giants v Washington Redskins Photo by Rob Carr/Getty Images

Pre-Competition Starting Roster

Putting all of the above together, the initial, partial starting roster, prior to the competition phase looks as follows. I have expanded the definition of starter slightly to include primary options at RB2/flex, Slot WR, Slot CB and KR/PR. Starting roster locks are indicated in bold:

QB Dwayne Haskins

RB1 Adrian Peterson

RB2/flex J.D. McKissic

LT ?

LG ?

C Chase Roullier

RG Brandon Scherff

RT Morgan Moses

WR1 Terry McLaurin

WR2 ?

WR slot Steven Sims

TE ?

DE Ryan Kerrigan

NT Daron Payne

DT One of Matt Ioannidis/Jon Allen

DE Montez Sweat

WLB ?

MLB John Bostic

SLB Cole Holcomb

CB1 Ronald Darby

CB2 Fabian Moreau

CB slot Kendall Fuller

FS Sean Davis

SS Landon Collins

KR/PR ?

K Dustin Hopkins

P Tress Way

LS Nick Sundberg


Competition for Starting Spots

The next phase of roster building is to simulate a competition amongst the rookies and first to third-year developmental players for the starting spots not occupied by locks. Young players who are themselves roster locks or inside edge players are excluded.

The first step was to compile a table listing the probabilities of each young player earning starting time based on their draft position using data compiled for a previous article. (I’ll get to UDFAs below). I then used the random number generation function in Excel to “roll the dice” to determine if the player earned a starting spot. This generates a random decimal between zero and one. If the “dice roll” output was less than or equal to a player’s starting probability, he earned a starting spot. For example, as the second overall pick in the draft, Chase Young’s probability of starting in his first year is 0.78. If the dice rolls 0 to 0.78 he earns a starting spot. If it comes up higher than 0.78, he doesn’t.

The data I used were actually the probabilities that players would earn eight or more starts in their first and third years in the league. To calculate starting probabilities for second-year players, I interpolated between the starting probabilities for first and third-year players at each pick number. When a player earned a starting spot, I then did another dice roll from zero to eight (random number x 8, rounded to whole numbers) to determine how many additional starts he earned, above the eight he started with when he made the cut, for a total of eight to 16 starts. The simulation didn’t cover players earning seven or fewer starts. Those may be filled in after the fact at my discretion as GM.

Determining starting probabilities for UDFAs was far more challenging. In large part, this is because it is very hard to keep track of how many UDFAs cycle through the training facility and roster every year. The initial priority UDFA signings that occur within the two to three weeks following the draft are well reported. But after that UDFAs come and go, often without leaving much record.

To get a handle on the starting hit rate of UDFAs, I focused on the priority UDFAs the Redskins have signed since 2010. In that period, the Redskins signed approximately 115 undrafted free agents immediately following the draft. Of those priority UDFAs, six (0.052) started at least one game as rookies and only one (0.009), Fat Rob Kelly, started more than eight games in his first year. Based on that fairly small sample size the probability of priority UDFAs earning starting time as rookies is roughly comparable to that of seventh round draft picks. If it seems that more UDFAs make the team than seventh round picks, that’s mainly because priority UDFAs outnumber seventh rounders by more than eight to one.

To make life simple, all of the UDFAs in this simulation were given the same probability of earning starting spots as the middle pick of round 7. In 2020 that was pick #230, which historically has a 0.03 probability of starting as a rookie. Their starting probabilities increase with years in the league just like seventh round draft picks.

Finally, in the last 10 years four UDFAs who signed later in the preseason earned some starting time as rookies. These UDFAs were TE Hale Hengtes (2019, 4 starts), CB Quinton Dunbar (2015, 2 starts), LB Jackson Jeffcoat (2014, 1 start), and OL Willie Smith (2011, 3 starts). Based on that history there is a 40% chance that a UDFA we haven’t seen yet makes the team and starts at least one game in 2020. Let’s just get that out of the way now and see if I have to come up with a mystery UDFA. Dice roll says 0.93 (0.93 > 0.40). No late season UDFA starter this year.

The starting probabilities for the young players competing for starting spots are listed in the following two tables. Remember that players competing for vacant/wide open starting spots get their probabilities doubled (last column). I’ll now reveal the results of the dice rolls. But bear in mind that most of the players who earned starting spots in 2020 won’t start immediately, and the final 53-man roster can’t be set until we determine preseason injuries and resolve conflicts.

In order to make sense of the results for myself and the reader, I made up a narrative to explain how the preseason played out according to the results of my simulations. This will continue through the second article as the regular season unfolds.

Drafted Players
Player Position Pick # p(start) Raw p(start) Adjusted
2020 Draft Class
Chase Young DE 1 0.78
Antonio Gibson RB 66 0.25
Saahdiq Charles OT 108 0.11 0.22
Antonio Gandy-Golden WR 142 0.07 0.14
Keith Ismael G/C 156 0.06 0.12
Khaleke Hudson LB 162 0.06 0.12
Kamren Curl S 216 0.04
James Smith-Williams DE 229 0.03
2019 Draft Class
Bryce Love RB 112 0.175
Wes Martin G 131 0.145 0.29
Ross Pierschbacher G/C 153 0.115 0.23
Kelvin Harmon WR 206 0.0625 0.125
Jimmy Moreland CB 227 0.045
Jordan Brailford DE 253 0.03
Caleb Wilson TE 254 0.03 0.06
2018 Draft Class
Derrius Guice RB 59 0.44
Geron Christian OT 74 0.36 0.72
Troy Apke S 109 0.25
Shaun Dion Hamilton LB 197 0.1 0.2
Greg Stroman CB 241 0.05
Trey Quinn WR 256 0.05 0.1

I won’t reveal the number of starts players earned in this article, just the names of the players who earned starting time, subject to injury. You may have guessed that the names in bold are the winners. There were no major surprises amongst the rookie draft class. Saahdiq Charles looked much better in training camp and preseason than a fourth round rookie. But he still struggled at times to pick up the blocking schemes which probably wasn’t helped by the impact of the lockdowns. I’d say he was neck and neck with Cornelius Lucas for the starting LT position, although they both had some unexpected competition from other players on the roster.

Chase Young looked as good as advertised in opponents’ backfields. However, his ability to pick up Del Rio’s scheme was clearly set back by lack of OTAs and minicamp, and he made plenty of rookie mistakes in preseason games.

There were quite a few surprises in the second year class. Bryce Love pulled off a stunning comeback from injury to make a case for time at RB2. However, he had strong competition for third-down back responsibilities from more versatile players J.D. McKissic and Antonio Gibson, plus another player we will get to a bit later.

In possibly the biggest surprise of the preseason, Ross Pierschbacher came to training camp in amazing shape and pulled even with classmate Wes Martin, whom many were expecting to compete with Wes Schweitzer for starting LG. Kelvin Harmon, who was looking good toward the end of the 2019 season, was seriously challenged for the starting WR2 position by rookie Antonio Gandy Golden. They were very close in preseason, and Harmon enters the season feeling pressure to retain his starting position.

Last season’s Mr. Irrelevant, TE Caleb Wilson was signed from the Cardinals’ practice squad to the Redskins’ active roster on December 13, 2019, where he made zero impact and did not play a single offensive snap. He must have put in a lot of work in the offseason, though. But his surprising emergence probably had as much to do with his dramatic improvement under the tutelage of master TE coach Pete Hoener as the absence of quality competition to start at TE. By the end of preseason it was not clear who would start between Wilson and FA acquisition Logan Thomas, with promising UDFA Thad Moss knocking on the door.

In the 2018 draft class, Derrius Guice came back strong from injury and once again looked like the best RB on the roster. Shaun Dion Hamilton also lived up to many fans’ expectations. The mild surprise in this group was Geron Christian, who looked pretty pedestrian in 2019. But Adrian Petersen and others have commented that he is looking much stronger this season and those reports were proven true. In case anyone is wondering if the doubling of probabilities for OTs gave Geron an unfair advantage, he actually didn’t need it. He rolled a 0.16, and would have earned a starting spot without the adjustment.

On to the first to third-year UDFAs. Their starting probabilities are listed in the following table. Names of lucky winners are again in bold.

Undrafted Free Agents
Player Position p(start) p(start) Adjusted
First Year UDFAs
Thaddeus Moss TE 0.03 0.06
Steven Montez QB 0.03
Isaiah Wright WR 0.03 0.06
Johnathon Johnson WR 0.03 0.06
Second Year UDFAs
Ryan Bee DT 0.05
Hale Hentges TE 0.05 0.1
Craig Reynolds RB 0.05
Jester Weah WR 0.05 0.1
Paul Adams OT 0.05 0.1
Emanuel Hall WR 0.05 0.1
Third Year UDFAs
Danny Johnson CB 0.07
Cam Sims WR 0.07 0.14
Darvin Kidsy WR 0.07 0.14
Timon Parris OT 0.07 0.14
Jeremy Reaves S 0.07
Kyle Allen QB 0.07
Jordan Veasy WR 0.07 0.14
Marcus Baugh TE 0.07 0.14

By now I can reveal that two of the big stories of the preseason were the emerging talents on the offensive line, and the rise of UDFA Isaiah Wright as the kind of versatile, offensive-playmaking threat that Scott Turner covets, showing great promise as a dual-threat WR/RB hybrid weapon, and establishing himself as the best punt returner on the team.

Perhaps the second-biggest surprise after Ross Pierschbacher was the dramatic improvement of Timon Parris, who was not far behind talented rookie Saahdiq Charles, in terms of blocking ability. Both were followed closely by ascendant third-year OT Geron Christian. Christian maintained one advantage over the other two, by showing greater mastery of the playbook and making no mistakes. What seemed like a gaping hole heading into camp turned into an embarrassment of riches, with four-way battle for one starting position. By the end of the preseason, Hogs Haven readers were proclaiming Carolina import OL coach John Matsko to be the true OL whisperer.

Mason Brennan Award Winner

There was a close race between Isaiah Wright, Timon Parris, and Caleb Wilson, early in training camp. By the second preseason game, it was down to Wright and Parris. But Wright pulled away from the other to take home the Mason Brennan award with a breakout performance in the third preseason game, making some clutch catches and ripping off some long punt returns, one for a score.

Reuben Foster

As I mentioned above, Reuben Foster is such a pivotal player that he deserves his own category. If healthy, he has elite potential and is the only starting quality player on the team with the skill set to play weakside linebacker in 4-3 defense. The next closest players, Josh Harvey-Clemons and Kevin Pierre-Louis have yet to demonstrate that they are NFL starters. After them, Cole Holcomb is not good in coverage, Thomas Davis is pushing 40, and Khaleke Hudson already missed out in the competition phase.

Foster is rehabbing from a severe knee injury with nerve damage, and caused such problems for the 49ers off the field that they released him with more than two years remaining on his rookie contract. If he does come back from injury, serious doubts remain about his ability to earn Ron Rivera’s trust as a player who fits the new team culture. I wouldn’t know where to begin to estimate his chances of overcoming both obstacles, so I’m just going to place each at 0.25 probability, resulting in a 0.0625 joint probability that both happen and Foster comes back to win the starting WLB position.

Dice roll say 0.0492. Unbelievable. Reuben Foster beats the odds to claim the starting WLB position. In doing so he also earns a nomination for NFL Comeback Player of the Year.


NFL-Houston Texans at Washington Redskins Photo by Jonathan Newton / The Washington Post via Getty Images

Injuries

The final step to establish the opening day roster is to determine which players suffered injuries in the preseason. The cumulative probabilities of the average NFL player missing any number of weeks to injury, from 0 to 16, calculated from the data presented in figure 1 of a Football Outsiders article by NFL injury expert Zach Binney, are presented in the Average Risk column of the following table.

Cumulative Probability
Weeks Missed Average Risk Iron Men High Risk
0 0.599 0.74 0.485
1 0.7 0.806 0.614
2 0.767 0.849 0.7
3 0.833 0.892 0.785
4 0.865 0.913 0.826
5 0.885 0.926 0.852
6 0.902 0.937 0.874
7 0.914 0.945 0.889
8 0.926 0.953 0.904
9 0.935 0.959 0.915
10 0.943 0.964 0.925
11 0.95 0.968 0.934
12 0.956 0.972 0.942
13 0.962 0.976 0.95
14 0.967 0.979 0.956
15 0.974 0.983 0.965
16 1.001 1 0.999

Not all players have average injury risk, however. In another study, Zach Binney found that 40% of players, overall, missed at least one week to injury each season. But among players who missed no time to injury in the preceding two years, that figure dropped to 26%. To reflect this reduced injury risk, I calculated a second set of cumulative probabilities for Iron Men, who missed no time to injury in the last two years (NFL or college). Here, the probability of escaping injury is increased to 0.74 and the probabilities for missing all numbers of weeks to injury are reduced proportionately.

If absence of a recent injury history reduces injury risk relative to the general player population, then it stands to reason that having a significant recent injury history should increase the risk, due to risk of reinjury, if nothing else. Therefore, I also calculated another set probabilities for “High Risk” players who have missed significant time to injury in the last two years. In the absence of a dataset to go on, I simply defined “significant time” as 10 or more weeks in two years. I made the “China Doll” effect equal and opposite to the Iron Man effect. Since the Iron Man effect increases the chance of escaping injury from 0.6 to 0.74, the China Doll effect reduces that chance by 0.6/0.74 to 0.48. All of the injury risks were then increased proportionately.

Despite the injury plague of recent years, nearly half (24) of the players who made the final 53 were Iron Men, having missed no games to injury in the last two years. Eighteen of the finalists were in the average risk group, having missed one to nine games. And 12 players were in the High Risk category, having missed 10 or more games.

To predict each player’s injury status for the season, I used the random number generator to “roll the dice”, yielding a random number from zero to one which was read off against the cumulative probabilities in the appropriate column of the table. For example, if a player who had missed three games to injury over the last two years rolled a 0.524, I look at the Average Risk column and see that 0.524 is below 0.599, so he escapes injury. If the same player had rolled 0.875, he would miss five weeks to injury.

After I determined which players missed time to injury, I then did a second dice roll for each injured player to determine when in the season the injury occurred from preseason to Week 16. For example, a player who missed five weeks to injury could go out anywhere from Week 1 (preseason injury) to Week 11. The injuries to other players will be revealed as the season plays out in the next article.

The following players had preseason injuries and missed making the opening day active roster:

RB Adrian Peterson – day-to-day nursing a hamstring injury, inactive for the season opener

NT Daron Payne - torn meniscus, possible return by late November to December

CB Aaron Colvin - torn achilles tendon, out for the season

OT Cornelius Lucas - torn ACL, out for the season


Final 53-Man Roster Selection

Now, I fit the competition winners into the final 53-man roster. None of the young new starters actually earned 16 starts via the dice rolls. Derrius Guice starts opening day at RB in place of the injured starter, Adrian Petersen. The rest of the competition winners take positions on the opening day roster behind last year’s starters, or veteran players that I slotted into vacancies using my discretion as GM.

The most difficult starting position to resolve was LT, where the leading veteran contender is out for season, and three younger players stepped up in preseason. Exercising my GM’s discretion, Geron Christian was slotted in as the day one starter, due to his greater experience and mistake-free play in preseason. The raw, but talented Charles makes the roster as his backup. As extra insurance, and to prevent the Colts or another team from poaching from the practice squad, the team makes the unusual move of keeping Parris as a fourth OT on the 53-man roster.

FA veteran Wes Schweitzer starts the season at LG. Pierschbacher joins draft classmate Wes Martin and 2020 fifth-round pick Keith Ismael as backup interior offensive linemen.

J.D. McKissic starts the season as the primary option in the third-down back/flex/weapon roll. He is joined in the RB room by comeback player Bryce Love and 2020 third round pick Antonio Gibson. Breakout player Isaiah Wright is primarily listed as a WR, but might be expected to take some snaps at RB, as well.

Kelvin Harmon retains his starting role at WR2 for the season opener. Antonio Gandy-Golden and Isaiah Wright earned positions in the WR lineup behind Harmon and returning veterans Terry McLaurin and Steven Sims. Cam Sims seems like the best available option for the sixth WR slot. Wright takes primary punt return duties on opening day. Steven Sims’ role on special teams is reduced to splitting time at KR with Antonio Gibson.

The open competition for starting TE did not turn up any clear-cut winners. Jeremy Sprinkle failed to impress in 13 starts last season and the new regime decides it’s time to move on. Rivera and Turner’s FA acquisition, Logan Thomas gets the nod to start. He is backed up by emerging talent Caleb Wilson and 2020 UDFA Thad Moss. Hale Hengtes is signed to the practice squad.


NFL: Washington Redskins at Philadelphia Eagles Eric Hartline-USA TODAY Sports

Final Opening Day Roster

The final 53 man roster for the season opener is as follows, with starters listed first:

QB Dwayne Haskins, Kyle Allen

RB1 Derrius Guice, (Adrian Peterson, inactive - injured)

RB2/flex J.D. McKissic, Bryce Love, Antonio Gibson

LT Geron Christian, Saahdiq Charles

LG Wes Schweitzer, Ross Pierschbacher

C Chase Roullier, Keith Ismael

RG Brandon Scherff, Wes Martin

RT Morgan Moses, Timon Parris

WR1 Terry McLaurin, Cam Sims

WR2 Kelvin Harmon, Antonio Gandy Golden

WR slot Steven Sims, Isaiah Wright

TE Logan Thomas, Caleb Wilson, Thad Moss

DE Ryan Kerrigan, Chase Young

NT Tim Settle, Caleb Brantley

DT Matt Ioannidis, Jonathan Allen,

DE Montez Sweat

WLB Reuben Foster, Thomas Davis Sr., Khaleke Hudson

MLB Jon Bostic, Shaun Dion-Hamilton

SLB Cole Holcomb, Ryan Anderson (also backup DE)

CB1 Ronald Darby, Greg Stroman

CB2 Fabian Moreau, Danny Johnson

CB slot Kendall Fuller, Jimmy Moreland

FS Sean Davis, Troy Apke

SS Landon Collins, Jeremy Reaves

KR Steven Sims, Antonio Gibson

PR Isaiah Wright

K/P/LS Tress Way, Dustin Hopkins, Nick Sundberg


Summary and Conclusions

The aim of this exercise was to generate a realistic opening day starting roster projection, which includes the types of surprises and unexpected developments which we see just about every year. To achieve this result, I introduced two elements of unpredictability into the normal roster projection process. I used NFL datasets to estimate the probabilities of (1) rookies and developmental players earning starting spots and (2) players getting injured throughout the season. Then I used a random number generator to roll the dice to generate one of many possible final starting roster scenarios.

The resulting opening day starting roster is pretty similar to what one would expect, but with a few surprises thrown in. Key starter Daron Payne was injured, but fortunately the team has solid depth at his position. Chase Young starts the season backing up and sharing time with Ryan Kerrigan. Somehow, I doubt it will take him long to earn the starting job, though. Derrius Guice, Bryce Love and Reuben Foster all came back from their injuries to make the roster.

Perhaps the biggest surprise of all was how the open competitions for OT and LG turned out. What looked like a massive problem area along the left side of the OL heading into camp was filled by several young players stepping up. Charles and Christian might have been expected, but Parris seems a bit less probable. Ross Pierschbacher surprised everyone by making a huge sophomore step to flash starting potential at LG, after looking pretty ordinary as a rookie.

A new, dynamic playmaking offensive weapon, Isaiah Wright, emerged from the rookie class, but probably not the one that most people expected (Gibson). Another unheralded player, Caleb Wilson, emerged from the developmental ranks at TE, but it remains unclear whether the team has found the long term solution at that position entering the season.

In the second installment of the series, we will see how the starting roster evolves as the 2020 season unfolds.

Poll

Which rookie not named Chase has the best chance to start this season?

This poll is closed

  • 16%
    Antonio Gibson
    (60 votes)
  • 60%
    Saahdiq Charles
    (222 votes)
  • 13%
    Antonio Gandy Golden
    (48 votes)
  • 7%
    Thad Moss
    (28 votes)
  • 0%
    Isaiah Wright
    (2 votes)
  • 2%
    Someone else
    (8 votes)
368 votes total Vote Now

Poll

Which developmental player breaks out in 2020?

This poll is closed

  • 37%
    Bryce Love
    (117 votes)
  • 33%
    Jimmy Moreland
    (104 votes)
  • 9%
    Hale Hengtes
    (29 votes)
  • 0%
    Paul Adams
    (1 vote)
  • 3%
    Timon Parris
    (10 votes)
  • 7%
    Cam Sims
    (22 votes)
  • 9%
    Someone else
    (28 votes)
311 votes total Vote Now

Poll

Which is most improbable this season?

This poll is closed

  • 11%
    Ross Pierschbacher claims a starting spot
    (36 votes)
  • 19%
    Timon Parris claims a starting spot
    (60 votes)
  • 3%
    Bryce Love coming back strong
    (10 votes)
  • 15%
    Derrius Guice playing a full season
    (47 votes)
  • 24%
    Reuben Foster healthy and on the roster in December
    (75 votes)
  • 15%
    Chase Young beat out for Redskins ROY
    (47 votes)
  • 10%
    Redskins miss out on the 2020 NFC championship
    (32 votes)
307 votes total Vote Now