Consistency Ratings? (1 Viewer)

[freshly_shorn]

With all of my leagues being head-to-head, it would be a valuable insight to have an 'at-a-glance' consistency rating for players' fantasy scoring. I believe this could be easily obtained by using the standard deviation of fantasy points scored per game. You can certainly get a feel by eye-balling game logs, but simply seeing an average points scored doesn't give you a feel for the boom-bust factor. I think it would be useful when trying to determine who to take between two players with the same apparent fantasy value.

I guess it would equate to a risk factor like the beta for stocks- how close to the his average are the player's actual week-to-week fantasy points?

Comments?

 

[gheemony]

Love the idea. Just add a column to the player stat pages. I think the next evolution is innovative stats on the stat pages.

 

[freshly_shorn]

Love the idea. Just add a column to the player stat pages. I think the next evolution is innovative stats on the stat pages.

Actually, I think customization is the next evolution. The MyFBG stuff is where you can get some real value-add for the price. Being able to present information in the context of your scoring system (and for multiple leagues) is great.But the consistency rating is an important consideration.. doesn't appear to send too many waves through the Pool, though.

 

[The Man with the Plan]

Consistency ratings are too "inconsistent" to count on from season to season.

 

[PahtyTom]

good idea, but consistency is overrated in fantasy football.

 

[AB in DC]

I believe this could be easily obtained by using the standard deviation of fantasy points scored per game.

That's fine for looking at past years' stats, but what makes you think that last year's SD has any relevance for projecting next year's stats?

 

[code]

Consistency has been discussed a number of times on these forums...the biggest problem is that the sample size (games in a single NFL season) is too small to derive much value from things like standard deviation and situations change enough over time (changes in coaches, schemes, teammates) that it is hard to build a good historical model...

 

[freshly_shorn]

I believe this could be easily obtained by using the standard deviation of fantasy points scored per game.

That's fine for looking at past years' stats, but what makes you think that last year's SD has any relevance for projecting next year's stats?

For that matter, what makes YOU think last year's stats have any relevance on future stats? They don't. But we use them to predict 'likelihoods' and tendencies, and I think SD can be used as just another nugget to see whether a player is a boom-or-bust type, or a consistent performer (compared to his average).

 

[Warpig]

Consistency ratings are too "inconsistent" to count on from season to season.

 

[freshly_shorn]

Consistency has been discussed a number of times on these forums...the biggest problem is that the sample size (games in a single NFL season) is too small to derive much value from things like standard deviation and situations change enough over time (changes in coaches, schemes, teammates) that it is hard to build a good historical model...

But you could make this argument against stats projections, too, but we do that anyway, right? So I don't see why you couldn't look at how those points have come historically. Keep in mind I'm not saying try to forecast consistency- I'm only interested in how consistent the player has been over time.

 

[Warpig]

With all of my leagues being head-to-head, it would be a valuable insight to have an 'at-a-glance' consistency rating for players' fantasy scoring. I believe this could be easily obtained by using the standard deviation of fantasy points scored per game. You can certainly get a feel by eye-balling game logs, but simply seeing an average points scored doesn't give you a feel for the boom-bust factor. I think it would be useful when trying to determine who to take between two players with the same apparent fantasy value.I guess it would equate to a risk factor like the beta for stocks- how close to the his average are the player's actual week-to-week fantasy points?Comments?

In my deeper leagues (40+ roster spots) when it comes to picking up players I always use my own consistency system. All too often you'll have people who say: "look at soandso, he scored 170 pts last year"...yeah, well, soandso scored 75 points in one game, 40 pts in another game, 25 in one game, and 30 in another, and big fat zeros for all the other games.I'll take the guy who, over the length of an entire season, scored 12 - 15 pts a game, regardless of which defense they were playing against. It's nice to know you can plug and go and know what you're getting from your lineup.

 

[bostonfred]

Help me understand how to use this information.

 

[freshly_shorn]

Help me understand how to use this information.

Warpig gives an example, but say Player A and B both average 12 points a game, but Player A gets numbers like 20. 3, 4, 25, etc. while Player B gets 10, 11, 14, 15, etc. throughout the season. Standard deviation for Player A would be much higher than for Player B, indicating the 'spread' around the average. I personally would rather have the steady and 'predictable' Player B over Player A.In other words, would it be useful to easily compare two 12-point per game players, knowing that they scored either 3-25 points per week or 10-15 points per week?

 

[code]

Help me understand how to use this information.

Warpig gives an example, but say Player A and B both average 12 points a game, but Player A gets numbers like 20. 3, 4, 25, etc. while Player B gets 10, 11, 14, 15, etc. throughout the season. Standard deviation for Player A would be much higher than for Player B, indicating the 'spread' around the average. I personally would rather have the steady and 'predictable' Player B over Player A.In other words, would it be useful to easily compare two 12-point per game players, knowing that they scored either 3-25 points per week or 10-15 points per week?

In other threads on this topic I've explained the one "consistency" statistic I've used in the past...(this, btw is where I developed the opinion that the sample size was too small to be meaningful)...I calculated the scoring average and standard deviation, subtracted the stdev from the average and called this the "base"...if the data had been a normal distribution I could've claimed that at least roughly 85% of the time the player would score at least the "base" score...but the data was nowhere near normally distributed and I eventually stopped tracking this stat altogether...

 

[GregR_2]

Consistency has been discussed a number of times on these forums...the biggest problem is that the sample size (games in a single NFL season) is too small to derive much value from things like standard deviation and situations change enough over time (changes in coaches, schemes, teammates) that it is hard to build a good historical model...

But you could make this argument against stats projections, too, but we do that anyway, right? So I don't see why you couldn't look at how those points have come historically. Keep in mind I'm not saying try to forecast consistency- I'm only interested in how consistent the player has been over time.

No, you can't make the same argument about stats projections because testing reality shows that last year's stats are a decent predictor of next year's stats. But doing the same statistical test with last year's consistency shows that it has no significant bearing on the future year's consistency.We're not just postulating ideas without testing them. That's the great thing about Footballguys. Doug Drinen who is a math professor has done studies to determine what the correlation between past and future consistency are. Back in 2000 he found that there were no significant amounts of correlation for QB, RB or WR consistency in scoring fantasy points from one year to the next. I believe I recall he did this again a few years later with the same result, though I didn't see a link for it when I did a quick google. We're not just pulling arguments out of our butts on this one. Consistency has been statistically tested and found that it isn't a good predictor for the following year. Believe me, if the evidence of the real world showed that it was a useful thing to include, the people on this board would have been all over using it years ago. If you can find evidence that the past results or wrong or that the league has changed, please share them because we'd love to see them. Until then, it's good to know where it's worth spending your time, isn't it?

 

[The Man with the Plan]

I believe this could be easily obtained by using the standard deviation of fantasy points scored per game.

That's fine for looking at past years' stats, but what makes you think that last year's SD has any relevance for projecting next year's stats?

For that matter, what makes YOU think last year's stats have any relevance on future stats? They don't. But we use them to predict 'likelihoods' and tendencies, and I think SD can be used as just another nugget to see whether a player is a boom-or-bust type, or a consistent performer (compared to his average).

That's not true. The best indicator of future performance is past performance. If you want to see how somebody will do in the future you look at what they did in the past. The problem with ranking how consistent a player is that you're trying to account for every little thing that led up to their fantasy point totals. It's like trying to count the sand in the desert. Also there's just no real correlation from year to year. One year a guys a consistent performer and a boom-or-bust player the next season. Case in point: Chad Johnson.Here are the point totals for Chad Johnson's last four seasons: 2003:195 2004:182 2005:200 2006:180In the 2003,2004, and 2005 seasons he was certainly considered a consistent player. Chad Johnson scored in double digits in 9 games in 2003, 7 games in 2004, and 9 games in 2005. In addition Chad Johnson only had 2 games in 03, 4 games in 04, and 1 game in 2005 in which he had under 50 yards receiving. Coming into the 2006 season there was no reason to think about Chad Johnson as anything else but a great wide receiver who was also a consistent scorer right?. Right?Fast forward to now and it seems that CJ is now labeled as an entirely inconsistent player. Just looking at his 2006 numbers how couldn't you? There was a 3 game stretch of greatness and not much else. What happened to old Mr. Consistent Chad Johnson? How and why did he accumulate the same point total (and the same position ranking) as he has in seasons past but in an entirely different way? If you can answer that question then perhaps you can determine if he'll be inconsistent in this season like he was last season or if he'll be as consistent as he was the 3 seasons previous. Honestly I don't know how to make that determination so that is why I can't put much stock in "consistency rankings". Maybe you can figure it out where I and others have failed.

 

[freshly_shorn]

Consistency has been discussed a number of times on these forums...the biggest problem is that the sample size (games in a single NFL season) is too small to derive much value from things like standard deviation and situations change enough over time (changes in coaches, schemes, teammates) that it is hard to build a good historical model...

But you could make this argument against stats projections, too, but we do that anyway, right? So I don't see why you couldn't look at how those points have come historically. Keep in mind I'm not saying try to forecast consistency- I'm only interested in how consistent the player has been over time.

No, you can't make the same argument about stats projections because testing reality shows that last year's stats are a decent predictor of next year's stats. But doing the same statistical test with last year's consistency shows that it has no significant bearing on the future year's consistency.We're not just postulating ideas without testing them. That's the great thing about Footballguys. Doug Drinen who is a math professor has done studies to determine what the correlation between past and future consistency are. Back in 2000 he found that there were no significant amounts of correlation for QB, RB or WR consistency in scoring fantasy points from one year to the next. I believe I recall he did this again a few years later with the same result, though I didn't see a link for it when I did a quick google. We're not just pulling arguments out of our butts on this one. Consistency has been statistically tested and found that it isn't a good predictor for the following year. Believe me, if the evidence of the real world showed that it was a useful thing to include, the people on this board would have been all over using it years ago. If you can find evidence that the past results or wrong or that the league has changed, please share them because we'd love to see them. Until then, it's good to know where it's worth spending your time, isn't it?

Very interesting, and well-put. Thanks! I, for one, would still like to see it, again- not as a predictor, but as another way to quantify past results. But, I can certainly do that on my own in the few instances I need to make this sort of decision. Thanks for the info, that certainly satisfies my curiosity as to why FBG doesn't do it.

 

[freshly_shorn]

I believe this could be easily obtained by using the standard deviation of fantasy points scored per game.

That's fine for looking at past years' stats, but what makes you think that last year's SD has any relevance for projecting next year's stats?

For that matter, what makes YOU think last year's stats have any relevance on future stats? They don't. But we use them to predict 'likelihoods' and tendencies, and I think SD can be used as just another nugget to see whether a player is a boom-or-bust type, or a consistent performer (compared to his average).

That's not true. The best indicator of future performance is past performance. If you want to see how somebody will do in the future you look at what they did in the past. The problem with ranking how consistent a player is that you're trying to account for every little thing that led up to their fantasy point totals. It's like trying to count the sand in the desert. Also there's just no real correlation from year to year. One year a guys a consistent performer and a boom-or-bust player the next season. Case in point: Chad Johnson.Here are the point totals for Chad Johnson's last four seasons: 2003:195 2004:182 2005:200 2006:180In the 2003,2004, and 2005 seasons he was certainly considered a consistent player. Chad Johnson scored in double digits in 9 games in 2003, 7 games in 2004, and 9 games in 2005. In addition Chad Johnson only had 2 games in 03, 4 games in 04, and 1 game in 2005 in which he had under 50 yards receiving. Coming into the 2006 season there was no reason to think about Chad Johnson as anything else but a great wide receiver who was also a consistent scorer right?. Right?Fast forward to now and it seems that CJ is now labeled as an entirely inconsistent player. Just looking at his 2006 numbers how couldn't you? There was a 3 game stretch of greatness and not much else. What happened to old Mr. Consistent Chad Johnson? How and why did he accumulate the same point total (and the same position ranking) as he has in seasons past but in an entirely different way? If you can answer that question then perhaps you can determine if he'll be inconsistent in this season like he was last season or if he'll be as consistent as he was the 3 seasons previous. Honestly I don't know how to make that determination so that is why I can't put much stock in "consistency rankings". Maybe you can figure it out where I and others have failed.

You are absolutely right- I should have mentioned (as I did later) that I was only interested in the past, not as a predictor. I agree- its futile to forecast consistency. But, I still hold that consistency in past results is a somewhat useful means of comparing otherwise equal players. The utility is, admittedly, limited, and I could certainly eyeball past results to get an idea.

 

[code]

Consistency has been discussed a number of times on these forums...the biggest problem is that the sample size (games in a single NFL season) is too small to derive much value from things like standard deviation and situations change enough over time (changes in coaches, schemes, teammates) that it is hard to build a good historical model...

But you could make this argument against stats projections, too, but we do that anyway, right? So I don't see why you couldn't look at how those points have come historically. Keep in mind I'm not saying try to forecast consistency- I'm only interested in how consistent the player has been over time.

No, you can't make the same argument about stats projections because testing reality shows that last year's stats are a decent predictor of next year's stats. But doing the same statistical test with last year's consistency shows that it has no significant bearing on the future year's consistency.We're not just postulating ideas without testing them. That's the great thing about Footballguys. Doug Drinen who is a math professor has done studies to determine what the correlation between past and future consistency are. Back in 2000 he found that there were no significant amounts of correlation for QB, RB or WR consistency in scoring fantasy points from one year to the next. I believe I recall he did this again a few years later with the same result, though I didn't see a link for it when I did a quick google. We're not just pulling arguments out of our butts on this one. Consistency has been statistically tested and found that it isn't a good predictor for the following year. Believe me, if the evidence of the real world showed that it was a useful thing to include, the people on this board would have been all over using it years ago. If you can find evidence that the past results or wrong or that the league has changed, please share them because we'd love to see them. Until then, it's good to know where it's worth spending your time, isn't it?

 

[bjabrad]

So the one thing that's consistent is change. Got it. This is useful and I've learned something here. But maybe this consistentcy quest isn't a waste of time after all if this info can add ammunition to point of view that most of us already believe. RB's are more valuable than WR's because they are more consistent.

If we can learn one thing from our apparently unfruitful quest toward identifying and drafting consistent FF players it's that WR's generally don't have the capability to be consistent. However, I think that common sense tells us that RB's do have the capability to be consistent. I know I'm just pointing out the obvious here but it's worth remembering.

RB's that aren't in RBBC are the lone ball carrier while there are at least two or three WR's, a TE or two as well as the RB catching passes. Does anybody really want any WR that plays for New England even if their name is Randy Moss? They just spread the ball around too much and it's the same reason that you need more RB's on your team than WR's and that you draft RB's higher than WR's.

When in doubt over a flex spot for example,,,play the RB if it's an option. I know,,,I know,,,but I'm just pointing out the painfully obvious here for the guppies and luckies in the shark pool.

 

[GregR_2]

So the one thing that's consistent is change. Got it. This is useful and I've learned something here. But maybe this consistentcy quest isn't a waste of time after all if this info can add ammunition to point of view that most of us already believe. RB's are more valuable than WR's because they are more consistent.

If we can learn one thing from our apparently unfruitful quest toward identifying and drafting consistent FF players it's that WR's generally don't have the capability to be consistent. However, I think that common sense tells us that RB's do have the capability to be consistent. I know I'm just pointing out the obvious here but it's worth remembering.

RB's that aren't in RBBC are the lone ball carrier while there are at least two or three WR's, a TE or two as well as the RB catching passes. Does anybody really want any WR that plays for New England even if their name is Randy Moss? They just spread the ball around too much and it's the same reason that you need more RB's on your team than WR's and that you draft RB's higher than WR's.

When in doubt over a flex spot for example,,,play the RB if it's an option. I know,,,I know,,,but I'm just pointing out the painfully obvious here for the guppies and luckies in the shark pool.

FBG David Shick! did an article about that called Standard Deviation. Couple of years old now but again, something I wouldn't expect to have changed much.I think the diagram near the bottom using the arrows along the lines is a telling one. His conclusions on that point:

So who’s more consistent? The WR’s have a lower standard deviation, but there’s a catch. The RB’s score more points than the WR’s do. If we talk about the average top 10 RB and top 10 WR, we need to consider how many points they are scoring to see an appropriate SPREAD of these points. Remember that the standard deviation is giving us a range above and below a player’s average. These are values that we expect to see about 68 percent of the time. Check out this graphic.

Neither position is way more consistent than the other is. The graphic above displays that the range for RB’s and WR’s are almost the same, but every now and then your top flight WR will score 5 points while your RB at least got you 10 points. It happens. It doesn’t mean that RB’s are more consistent. It means they score more points!

 

[phthalatemagic]

I believe this could be easily obtained by using the standard deviation of fantasy points scored per game.

Yes, but you'd have to relate each game to the opposing defense. If you want to be more accurate, then you'd have to rate the particular defense as they were playing during that time of year. You might get into some subjective numbers, and I don't know how you can resolve that. Even after you could figure all that out, that doesn't mean the player will be consistent the next year, it just says that particular player was consistent the year before. Perhaps if you do a three year study on players and see particular players tend to be more consistent, you could see something worthwhile. I'm betting players will be all over the place though, and the numbers will be inconclusive.I like the idea though, and I factor that in to my own rankings. I don't weigh it very heavily.

 

[phthalatemagic]

Until then, it's good to know where it's worth spending your time, isn't it?

Sometimes wisdom is best learned the hard way.