What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Week to Week Player Consistency/Volatility Analysis (1 Viewer)

Scoresman

Footballguy
One of the most annoying things to deal with in Fantasy Football is making lineup decisions and dealing with guys who fluctuate between having great weeks and bad weeks. We generally draft based rankings which are based on total point projections, but those projections rely on you making the correct start/sit decision with those players, something that is anything but an exact science for anyone other than the studs.

The poster child for this last year was Gabe Davis. He had a few good/great games where he flirted with 20 PPG, but he had 7, SEVEN games where he scored less than 1 point. That's nuts and probably made him a wasted draft pick if you didn't correctly predict all those goose eggs.

One analysis I'm playing with this year is an assessment of each player identifying their volatility, putting it on my cheat sheet and avoiding highly volatile players as much as possible.

How do I measure volatility?

I started with standard deviations from position weekly averages. I started here because standard deviation sounds like it would measure exactly what we need, the deviation a set of weekly scores has from the mean. This ended up not working. As an example, Josh Allen showed up as a highly volatile player because he had games where he scored 40 points, and games where he scored 18-20 points. This is volatile, but to be honest, I'm fairly happy with anything within that range of points for my starting QB and it wouldn't be a reason not to draft Allen.

What I really wanted to measure was "How many games did each player have where I would have regretted adding them to my starting lineup in hindsight?" This led me to going through each player's weekly scores and tallying which weeks were unacceptable, and then coming up with a percentage of "startable" weeks for each player. To use the Josh Allen example, he ended up with 1 "unacceptable" start, a 12 point performance in Week 1, so he gets a 93% startable rate. Obviously with Josh Allen, you'll be starting him every week regardless but it's nice to see how other players compare.

This is a bit subjective as the "acceptable" number of points a player gives is different for each position and tier. When you start Justin Jefferson, you expect a certain minimum number of points, and that number of points is probably not the same as if you were starting Romeo Doubs. I'm still tweaking how I calculate all this, but this is what I have so far. It also only has a sample size of one season so far.

TLDR: The higher the percentage, the higher number of weeks you should have felt good having him in your starting lineup.

6/17 Update:

Tables below have been updated.

I went through everything and stacked all of the weekly points with weekly snap counts because I realized I had a big problem of including games where players left early due to injury or were just really banged up and played significantly fewer snaps than their average. I went through and removed these games from their sample sizes. So a couple reminders:

- The percentages represent the percentage of weeks where a player provided an "acceptable" fantasy performance. In other words, it's the percent of weeks you would not have regretted starting them.
- Does not include games with limited snaps due to injury.
- It's a sliding scale. Christian McAffrey putting up 8 points in a week is a disappointment, but Hunter Renfrow putting up 8 points is pretty good.
- This is still not an exact science and I'm still unsure how much I will use this. Players changing teams, having their coaching staff change, or many other factors can affect volatility.
 
Last edited:
NAMEPOSTeamPercent
Josh AllenQBBUF
94%​
Kyler MurrayQBAZ
93%​
Baker MayfieldQBTB
91%​
Jordan LoveQBGB
88%​
Matthew StaffordQBLAR
86%​
Patrick Mahomes IIQBKC
84%​
C.J. StroudQBHOU
82%​
Brock PurdyQBSF
81%​
Jalen HurtsQBPHI
81%​
Justin HerbertQBLAC
81%​
Russell WilsonQBDEN
80%​
Trevor LawrenceQBJAX
80%​
Sam HowellQBWAS
79%​
Dak PrescottQBDAL
78%​
Jared GoffQBDET
78%​
Lamar JacksonQBBAL
78%​
Tua TagovailoaQBMIA
78%​
Derek CarrQBNO
75%​
Kirk CousinsQBMIN
75%​
Justin FieldsQBCHI
73%​
Geno SmithQBSEA
64%​
Joe BurrowQBCIN
60%​
Desmond RidderQBATL
50%​
Kenny PickettQBPIT
45%​
Ryan TannehillQBTEN
43%​
Bryce YoungQBCAR
40%​


Up next: RBs
 
Last edited:
Running Backs

NAMEPOSTeamPercent
David MontgomeryRBDET
92%​
Kyren WilliamsRBLAR
92%​
Christian McCaffreyRBSF
91%​
Jonathan TaylorRBIND
88%​
Raheem MostertRBMIA
88%​
D'Andre SwiftRBPHI
82%​
Kenneth Walker IIIRBSEA
80%​
Rhamondre StevensonRBNE
80%​
Zack MossRBIND
80%​
Chuba HubbardRBCAR
78%​
Jerome FordRBCLE
78%​
Bijan RobinsonRBATL
77%​
Alvin KamaraRBNO
75%​
Isiah PachecoRBKC
75%​
Rachaad WhiteRBTB
75%​
Derrick HenryRBTEN
73%​
James ConnerRBAZ
73%​
Joe MixonRBCIN
72%​
Brian Robinson Jr.RBWAS
71%​
AJ DillonRBGB
69%​
Devin SingletaryRBHOU
69%​
Josh JacobsRBLV
69%​
De'von AchaneRBMIA
69%​
Jahmyr GibbsRBDET
67%​
Khalil HerbertRBCHI
67%​
Saquon BarkleyRBNYG
67%​
Breece HallRBNYJ
64%​
Travis Etienne Jr.RBJAX
63%​
Javonte WilliamsRBDEN
62%​
James CookRBBUF
61%​
Antonio GibsonRBWAS
60%​
Najee HarrisRBPIT
60%​
Dameon PierceRBHOU
57%​
Gus EdwardsRBBAL
57%​
Jaylen WarrenRBPIT
56%​
Ezekiel ElliottRBNE
53%​
Austin EkelerRBLAC
50%​
Cam AkersRBLAR
50%​
Tony PollardRBDAL
50%​
Samaje PerineRBDEN
44%​
Zach CharbonnetRBSEA
44%​
Aaron JonesRBGB
40%​
Alexander MattisonRBMIN
38%​

Interesting seeing some highly touted backs rank so low here.
 
Last edited:
Wide Receivers.

After doing this exercise I realized that my primary receivers last year of Ridley, Lockett, and Gabe Davis were the reason I lost in my auction league. :lmao:


NAMEPOSTeamPercent
Amon-Ra St. BrownWRDET
93%​
Rashee RiceWRKC
87%​
CeeDee LambWRDAL
84%​
Mike EvansWRTB
83%​
Courtland SuttonWRDEN
82%​
Christian KirkWRJAX
82%​
Jaylen WaddleWRMIA
81%​
Keenan AllenWRLAC
81%​
Brandon AiyukWRSF
79%​
Chris OlaveWRNO
77%​
Tyreek HillWRMIA
77%​
A.J. BrownWRPHI
75%​
Michael Pittman Jr.WRIND
75%​
Jayden ReedWRGB
73%​
Nico CollinsWRHOU
73%​
Adam ThielenWRCAR
72%​
Puka NacuaWRLAR
72%​
DK MetcalfWRSEA
70%​
Deebo Samuel Sr.WRSF
69%​
George PickensWRPIT
69%​
Jordan AddisonWRMIN
64%​
Robert WoodsWRHOU
64%​
Tyler BoydWRCIN
64%​
Jakobi MeyersWRLV
63%​
Chris GodwinWRTB
63%​
Justin JeffersonWRMIN
63%​
Stefon DiggsWRBUF
63%​
Brandin CooksWRDAL
60%​
Diontae JohnsonWRPIT
60%​
Ja'Marr ChaseWRCIN
60%​
Justin WatsonWRKC
60%​
Tee HigginsWRCIN
60%​
Garrett WilsonWRNYJ
59%​
Zay FlowersWRBAL
59%​
Jameson WilliamsWRDET
59%​
Curtis SamuelWRWAS
58%​
Jerry JeudyWRDEN
57%​
Deandre HopkinsWRTEN
56%​
DeVonta SmithWRPHI
56%​
DJ MooreWRCHI
56%​
Elijah MooreWRCLE
56%​
Jaxon Smith-NjigbaWRSEA
56%​
Terry McLaurinWRWAS
56%​
Michael ThomasWRNO
56%​
Cooper KuppWRLAR
55%​
Hollywood BrownWRAZ
54%​
Amari CooperWRCLE
53%​
Romeo DoubsWRGB
53%​
Tyler LockettWRSEA
53%​
Christian WatsonWRGB
50%​
Davante AdamsWRLV
50%​
Drake LondonWRATL
50%​
Gabe DavisWRBUF
47%​
Calvin RidleyWRJAX
44%​
Jahan DotsonWRWAS
44%​
Odell Beckham Jr.WRBAL
44%​
Darnell MooneyWRCHI
36%​
DJ Chark Jr.WRCAR
36%​
Quentin JohnstonWRLAC
36%​
Rashod BatemanWRBAL
36%​
Zay JonesWRJAX
33%​
Michael GallupWRDAL
31%​
DeVante ParkerWRNE
30%​
Skyy MooreWRKC
25%​
Allen LazardWRNYJ
23%​
Juju Smith-SchusterWRNE
22%​
Treylon BurksWRTEN
22%​
Hunter RenfrowWRLV
18%​
Marquez Valdes-ScantlingWRKC
6%​
 
Last edited:
I've done something similar in the past, but instead of an arbitrary "acceptable" points, I calculated how many weeks the players was a "starter" at the position.

Assuming a 12 team, 1 QB league, I'd calculate how often a player was a top 12 QB in a given week. Do the same for the other positions based on how many starters by position. I'd even play with my version of arbitrary and do something such as look at top 4 weeks by QB, or top 12 weeks for RB/WR to identify "elite" weeks.
 
Are you including games that they were injured as "unacceptable" or not? Either way, something doesn't add up.

I believe AJ played all 17 and Devonta played in 16 but both are listed at 83%. 14/17 would be 82% while 13/16 would be 81%.

Apologies if its just an error, don't mean to nitpick.
 
Are you including games that they were injured as "unacceptable" or not? Either way, something doesn't add up.

I believe AJ played all 17 and Devonta played in 16 but both are listed at 83%. 14/17 would be 82% while 13/16 would be 81%.

Apologies if its just an error, don't mean to nitpick.

I dont inlude missed games due to injury or games that were cut short due to an injury (may not have gotten all of these but got most).

Also not including week 17 since its so screwy with guys not playing.
 
I've done something similar in the past, but instead of an arbitrary "acceptable" points, I calculated how many weeks the players was a "starter" at the position.

Assuming a 12 team, 1 QB league, I'd calculate how often a player was a top 12 QB in a given week. Do the same for the other positions based on how many starters by position. I'd even play with my version of arbitrary and do something such as look at top 4 weeks by QB, or top 12 weeks for RB/WR to identify "elite" weeks.

I thought about doing it this way but struggled to find week to week stuff that was easily downloadable. I'm thinking it'd come out very similar to mine. Mine's probably a bit tougher on what is considered "acceptable". My point in doing this was to weed out players who have tons of disastrous weeks among their good weeks.
 
Russel wilson with a surprisingly high score here. He had a lot of average weeks, not many duds, and almost no amazing weeks
Heard someone say a few weeks ago that Russell Wilson had more top 12 weeks last year then Stroud. I looked it up but only comped weeks they both played and Stroud won 6 to 5 but still speaks to how surprisingly consistent Wilson.

Which one one of two surprising Wilson stats I noticed, the other one being he led or tied for the NFL lead in 4th quarter comebacks.
 
Are you including games that they were injured as "unacceptable" or not? Either way, something doesn't add up.

I believe AJ played all 17 and Devonta played in 16 but both are listed at 83%. 14/17 would be 82% while 13/16 would be 81%.

Apologies if its just an error, don't mean to nitpick.

I dont inlude missed games due to injury or games that were cut short due to an injury (may not have gotten all of these but got most).

Also not including week 17 since its so screwy with guys not playing.
I like what you did here, good work, but for sure would have liked it more if we could have weeded out stuff like this but understand the limitations on the data.

Really throws a guy like Achane off when he early exited two of his 11 games.
 
Updated the running back table after finding more games where a player left with injuries. That's the hardest thing to weed out. There's nowhere I could find that tracks when a player left a game with an injury. Missing entire games is easy to spot and weed out, but its those games where a player plays, but leaves and doesn't return that I'm struggling with.

If anyone thinks a player's percent is off, let me know because it might be due to this.
 
Tables above have been updated.

I went through everything and stacked all of the weekly points with weekly snap counts because I realized I had a big problem of including games where players left early due to injury or were just really banged up and played significantly fewer snaps than their average. I went through and removed these games from their sample sizes. So a couple reminders:

- The percentages represent the percentage of weeks where a player provided an "acceptable" fantasy performance. In other words, it's the percent of weeks you would not have regretted starting them.
- Does not include games with limited snaps due to injury.
- It's a sliding scale. Christian McAffrey putting up 8 points in a week is a disappointment, but Hunter Renfrow putting up 8 points is pretty good.
- This is still not an exact science and I'm still unsure how much I will use this. Players changing teams, having their coaching staff change, or many other factors can affect volatility.
 
The percentages represent the percentage of weeks where a player provided an "acceptable" fantasy performance. In other words, it's the percent of weeks you would not have regretted starting them.
This should be based on scoring "starter" quality points. top 12 QB in start one QB for example. It shouldn't be a sliding scale because then you are hitting a moving target. Another piece to the puzzle is "game winning" weeks and "bust" weeks. I find it best to use three categories for this kind of evaluation because all of that info is important. Essentially trying to figure out average point totals for you scoring of the QB12, RB24, WR36, etc (considering a 1QB/2RB/3WR start league) and using that as a standard number for comparison. You can use historical data from your league if you have it and come up with an X-yr average (pick whatever you can evaluate - 5 yr average is probably good enough).

Then you can come up with "winning" weeks and figure out the QB3 average weekly score and use that as a threshold. Same with "bust". Could use 50% of the starter baseline numbers (or whatever percentage you would consider a bust week).

Use this as some baselines and it should be a pretty good metric for overall consistency/volatility.

ETA: Removing the low snap count weeks kind of skews the numbers as well. If a player is repeatedly not finishing games that should be included in their consistency/volatility ratings. You shouldn't just exclude that info because it does have an impact to the "startability" info you are mining.
 
Last edited:
avoiding highly volatile players as much as possible.
I don't think this is necessarily a good approach. I think high volatility players that boom to the point of week winning performances has value but as part of your roster construction. This is much more easily accomplished on the IDP side of things because the nature of IDP scoring lends itself to a lot of these type players. But it can be beneficial on the offensive side too.

Now, I don't want Boom/Bust guys as my cornerstone players but I don want a couple on my team and in use them in my lineups at times. Of course, you have to have an idea as to what type of matchups/situations lead to the boom or bust weeks so you can have the best chance of hitting the right way on those. But in games where I am a big underdog I like to have these type of guys in the lineup because them hitting gives me a big leg up on an opponent that should outscore me if averages play out.

All that to say, I don't think these type players are avoids. I think they have a purpose and a worth rostering for the right price and roster construction.
 
The percentages represent the percentage of weeks where a player provided an "acceptable" fantasy performance. In other words, it's the percent of weeks you would not have regretted starting them.
This should be based on scoring "starter" quality points. top 12 QB in start one QB for example. It shouldn't be a sliding scale because then you are hitting a moving target. Another piece to the puzzle is "game winning" weeks and "bust" weeks. I find it best to use three categories for this kind of evaluation because all of that info is important. Essentially trying to figure out average point totals for you scoring of the QB12, RB24, WR36, etc (considering a 1QB/2RB/3WR start league) and using that as a standard number for comparison. You can use historical data from your league if you have it and come up with an X-yr average (pick whatever you can evaluate - 5 yr average is probably good enough).

Then you can come up with "winning" weeks and figure out the QB3 average weekly score and use that as a threshold. Same with "bust". Could use 50% of the starter baseline numbers (or whatever percentage you would consider a bust week).

Use this as some baselines and it should be a pretty good metric for overall consistency/volatility.

ETA: Removing the low snap count weeks kind of skews the numbers as well. If a player is repeatedly not finishing games that should be included in their consistency/volatility ratings. You shouldn't just exclude that info because it does have an impact to the "startability" info you are mining.

I did this analysis really for just one reason, to try and reduce the problem of starting a player and being disappointed with the results. As a Gabe Davis and Tyler Lockett owner last year, this was incredibly frustrating. Game winning weeks is definitely another useful feature, but not what I'm doing right now. I'm trying to draft more consistent players in the mid-late rounds, not the boom or bust guys.

I essentially did what you describe by tiers. There has to be a sliding scale from tier to tier. If you're starting Tyreek Hill as your WR1, there's a certain minimum number of points you will be satisfied with and that's a different number than if you're starting Gabe Davis in your flex. When I'm looking at guys to draft in the mid-late rounds, I'm going to want to avoid the volatile guys because it requires you making the correct lineup decisions. This tool will help me draft more consistent players in those spots.

Also, I use a separate injury history metric to identify players who get banged up a lot.

I don't think this is necessarily a good approach. I think high volatility players that boom to the point of week winning performances has value but as part of your roster construction. This is much more easily accomplished on the IDP side of things because the nature of IDP scoring lends itself to a lot of these type players. But it can be beneficial on the offensive side too.

Now, I don't want Boom/Bust guys as my cornerstone players but I don want a couple on my team and in use them in my lineups at times. Of course, you have to have an idea as to what type of matchups/situations lead to the boom or bust weeks so you can have the best chance of hitting the right way on those. But in games where I am a big underdog I like to have these type of guys in the lineup because them hitting gives me a big leg up on an opponent that should outscore me if averages play out.

All that to say, I don't think these type players are avoids. I think they have a purpose and a worth rostering for the right price and roster construction.

I had Gabe Davis/Ridley/Lockett as rotating WR's in my WR2 slot. It was a nightmare trying to come up with who to start. I did PLENTY of research on matchups and it still just felt totally random most of the time. Believe me, this type of player is an AVOID. My team would have been a lot better drafting a more consistent player than these guys because those boom weeks require you to make the right decision to start them. Roster construction is easy on draft day, until you get to week 1 and have to decide who to start.

That's really why I did this analysis. To try and highlight the mid-late round guys who will be more consistent than the pack. I don't envision using this at all for studs who you will always start regardless.
 
I keep mentioning him but Gabe Davis last year is the poster boy for this analysis. Here are his weekly snaps stacked with his weekly points.

311​
Gabe DavisSnapsWRBUF
65​
58​
52​
42​
54​
54​
68​
68​
54​
53​
67​
91​
bye
63​
49​
47​
63​
312​
Gabe DavisPointsWRBUF
3.2​
15.2​
9.3​
12.1​
16​
0.1​
0.6​
14.7​
0​
7.6​
0​
16.5​
BYE
0​
0​
19​
2.1​

Snaps were fairly consistent, but look at those points. 49 snaps one week with zero points, 47 snaps the next week with 19 points? Or earlier in the year, two back to back weeks with 68 snaps, one week with 0.6 points, the other, 14.7 points.

The problem with a player like this is you get that one zero week and you're probably not going to want to start him the week after where he booms. That's why I'm not looking at the booms too much. You have to start them to get those points. I'd rather have consistent scorers for my non-stud players.
 
I'm trying to draft more consistent players in the mid-late rounds, not the boom or bust guys.
I actually think of it the opposite. I don't mind the boom/bust guys in the mid to late rounds because you aren't getting a "starter" quality guy most of the time so I want guys that can boom if I took care of my business on the front end of the draft with high consistency players. I put my "risk" in later round guys so having boom potential is a key positive for me.

Granted, you don't want guys that score zero 90% of the time with a 10% boom but drafting for upside late is what I try and do.
 
Thanks for doing this @Scoresman. Well done. We've been kicking around something similar and have some of the same questions about usability. Not sure if it's something we'll do but just wrestling with it can be useful. Thanks for the discussion.
 
I'm trying to draft more consistent players in the mid-late rounds, not the boom or bust guys.
I actually think of it the opposite. I don't mind the boom/bust guys in the mid to late rounds because you aren't getting a "starter" quality guy most of the time so I want guys that can boom if I took care of my business on the front end of the draft with high consistency players. I put my "risk" in later round guys so having boom potential is a key positive for me.

Granted, you don't want guys that score zero 90% of the time with a 10% boom but drafting for upside late is what I try and do.

I think you can draft for upside but also limit weekly volatility.
 

Users who are viewing this thread

Top