Strength of Schedule (1 Viewer)

[Mr. Sparkle]

Is there an article on strength of schedule on FBG.com ?

I checked, but did not see one. Thanks.

 

[Doug Drinen]

My guess is that it will appear soon. But I'm not 100% sure. Clayton Gray is the guy to ask. Shoot him a PM.

 

[purestrength]

Isn't there usually a helpful graphic that tells average FF points per position for each team's schedule as well?

 

[Bri]

Carlton

 

[Clayton Gray]

It will be finished in the next 1-3 days.

 

[(HULK)]

It will be finished in the next 1-3 days.

 

[AB in DC]

What I'd really like to see is a retroactive SOS analysis. It's very hard to predict opponents' performance for the upcoming season, but it's easy to look back at the previous season, see who faced easier or harder defenses vs. their position, and adjust accordingly.

 

[Bri]

It will be finished in the next 1-3 days.

Carlton,sorta like what Abrecher said. Do you notice anything year to year when doing this? If so what?

 

[bigreese82]

It will be finished in the next 1-3 days.

Carlton,sorta like what Abrecher said. Do you notice anything year to year when doing this? If so what?

I play in a co-owner league and my partner gets carried away IMO...with SOS...any trends or does it work better for different positions?

Would love to see a breakdown of whether SOS affected a players PPG avg?...I think I worded that right

 

[Chase Stuart]

What I'd really like to see is a retroactive SOS analysis. It's very hard to predict opponents' performance for the upcoming season, but it's easy to look back at the previous season, see who faced easier or harder defenses vs. their position, and adjust accordingly.

I think you mean something like this.I'll be writing a new and improved version in the coming weeks. I've given a lot of thought about expanding this to other positions, but that's proven pretty difficult. I might give it a shot anyway, but I'll be most confident about the QB one. Very interesting results last year indeed. This year the numbers aren't quite as revealing (i.e., adjusting 2005 fantasy totals for SOS doesn't change the rankings much) but that doesn't detract from the quality and purpose of the exercise.

If this is what you mean, throw out some suggestions on how to improve it. If not, ignore everything I just wrote.

 

[AB in DC]

What I'd really like to see is a retroactive SOS analysis. It's very hard to predict opponents' performance for the upcoming season, but it's easy to look back at the previous season, see who faced easier or harder defenses vs. their position, and adjust accordingly.

I think you mean something like this.

Basically, yes, but for all positions. I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

 

[Chase Stuart]

What I'd really like to see is a retroactive SOS analysis. It's very hard to predict opponents' performance for the upcoming season, but it's easy to look back at the previous season, see who faced easier or harder defenses vs. their position, and adjust accordingly.

I think you mean something like this.

Basically, yes, but for all positions. I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

It's a bit complicated but the main reason is there's only one QB on any given play. The nature of football makes analyzing RBs, WRs and TEs a bit more difficult when it comes to rear view SOS. I know a few other people feel the same way, so feel free to give us any suggestions on how you think the numbers should be run for those positions. I'm definitely open to suggestions and there's a decent chance I'll have an article on this at some point in the summer.

 

[AB in DC]

Basically, yes, but for all positions. I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

It's a bit complicated but the main reason is there's only one QB on any given play. The nature of football makes analyzing RBs, WRs and TEs a bit more difficult when it comes to rear view SOS. I know a few other people feel the same way, so feel free to give us any suggestions on how you think the numbers should be run for those positions. I'm definitely open to suggestions and there's a decent chance I'll have an article on this at some point in the summer.

Just use percentages.Suppose that Pittsburg's opponents averaged 15.0 PPG to RBs, vs. a leaguewide average of 16.5 PPG. (I'm making those numbers up.) In looking at Willie Parker's performance, I'd want to multiply his actual PPG by 16.5/15.0, or 110%, to get an adjusted PPG for 2005.

For those of us who use 2005 stats as the baseline for 2006 projections (as I do), I'd want to do this for every RB to see which players weren't really as good as their numbers would indicate, and which players were a lot better.

In this hypothetical case, Wilile Parker was really 10% better than his numbers due to a tough schedule.

 

[Chase Stuart]

Basically, yes, but for all positions. I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

It's a bit complicated but the main reason is there's only one QB on any given play. The nature of football makes analyzing RBs, WRs and TEs a bit more difficult when it comes to rear view SOS. I know a few other people feel the same way, so feel free to give us any suggestions on how you think the numbers should be run for those positions. I'm definitely open to suggestions and there's a decent chance I'll have an article on this at some point in the summer.

Just use percentages.Suppose that Pittsburg's opponents averaged 15.0 PPG to RBs, vs. a leaguewide average of 16.5 PPG. (I'm making those numbers up.) In looking at Willie Parker's performance, I'd want to multiply his actual PPG by 16.5/15.0, or 110%, to get an adjusted PPG for 2005.

For those of us who use 2005 stats as the baseline for 2006 projections (as I do), I'd want to do this for every RB to see which players weren't really as good as their numbers would indicate, and which players were a lot better.

In this hypothetical case, Wilile Parker was really 10% better than his numbers due to a tough schedule.

This might work.For Parker it's not too bad, but look at weeks 5 and 9. He had only 15 total carries those two weeks; Bettis had 17 carries in week 5, Staley had 15 in week 9.

Parker's opponents those weeks were GB and SD. GB was exactly average last year in FP allowed to RBs, while SD was pretty stingy. So do we credit Parker a whole game for each of those weeks? A half game? How does that work out.

With QBs, we can very accurately state that Carson Palmer played 15.1 games last year, that Chris Simms played 10.4 games, and that Marc Bulger played 6.8. With RBs it is a bit dicier. I think it gets even worse with WRs.

I'll re-think this one over and get back to you. I like your suggestion though and maybe that's the way to go. Either way I still like the idea much better for QBs.

 

[Bri]

Basically, yes, but for all positions.  I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

It's a bit complicated but the main reason is there's only one QB on any given play. The nature of football makes analyzing RBs, WRs and TEs a bit more difficult when it comes to rear view SOS. I know a few other people feel the same way, so feel free to give us any suggestions on how you think the numbers should be run for those positions. I'm definitely open to suggestions and there's a decent chance I'll have an article on this at some point in the summer.

Just use percentages.Suppose that Pittsburg's opponents averaged 15.0 PPG to RBs, vs. a leaguewide average of 16.5 PPG. (I'm making those numbers up.) In looking at Willie Parker's performance, I'd want to multiply his actual PPG by 16.5/15.0, or 110%, to get an adjusted PPG for 2005.

For those of us who use 2005 stats as the baseline for 2006 projections (as I do), I'd want to do this for every RB to see which players weren't really as good as their numbers would indicate, and which players were a lot better.

In this hypothetical case, Wilile Parker was really 10% better than his numbers due to a tough schedule.

I think your math is wrong and it should be lower. 1.5 less than avg not 110% or more than average

 

[Chase Stuart]

Basically, yes, but for all positions. I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

It's a bit complicated but the main reason is there's only one QB on any given play. The nature of football makes analyzing RBs, WRs and TEs a bit more difficult when it comes to rear view SOS. I know a few other people feel the same way, so feel free to give us any suggestions on how you think the numbers should be run for those positions. I'm definitely open to suggestions and there's a decent chance I'll have an article on this at some point in the summer.

Just use percentages.Suppose that Pittsburg's opponents averaged 15.0 PPG to RBs, vs. a leaguewide average of 16.5 PPG. (I'm making those numbers up.) In looking at Willie Parker's performance, I'd want to multiply his actual PPG by 16.5/15.0, or 110%, to get an adjusted PPG for 2005.

For those of us who use 2005 stats as the baseline for 2006 projections (as I do), I'd want to do this for every RB to see which players weren't really as good as their numbers would indicate, and which players were a lot better.

In this hypothetical case, Wilile Parker was really 10% better than his numbers due to a tough schedule.

I think your math is wrong and it should be lower. 1.5 less than avg not 110% or more than average

The way he did it was right. It can get confusing when you're talking about points allowed. But if the 16 teams on Team X's schedule allowed 15 PPG to RBs, and the league average allowed to RBs is 16.5, then RB Y faced a schedule where the average team RB would score 15 PPG, not 16.5 PPG (the average team RB by definition must score 16.5 PPG). Therefore you would want to increase his total by 110% (or 1.5 FPs, which I'd prefer I think).

 

[AB in DC]

For Parker it's not too bad, but look at weeks 5 and 9. He had only 15 total carries those two weeks; Bettis had 17 carries in week 5, Staley had 15 in week 9. Parker's opponents those weeks were GB and SD. GB was exactly average last year in FP allowed to RBs, while SD was pretty stingy. So do we credit Parker a whole game for each of those weeks? A half game? How does that work out.

Treat them as one whole game. There's no need to subdivide it any further. Do you adjust your QB stats when a QB throws 15 passes one game and 35 passes the next? The fact that Parker is in a RBBC is already accounted for in his PPG average. I suppose you could adjust based on the number of carries (similar to how FootballOutsiders do their stats), but you're looking at his game performance. If he's up against a tough defense and only has six carries, then so be it. You don't adjust the San Diego's Points Allowed stats just because one team ran 10 times against them and another team ran 25 times. Points Allowed is Points Allowed.

With QBs, we can very accurately state that Carson Palmer played 15.1 games last year, that Chris Simms played 10.4 games, and that Marc Bulger played 6.8. With RBs it is a bit dicier. I think it gets even worse with WRs.

I don't like to do fractional games, if at all possible. If he played, it's a whole game. If a player misses a half due to injury, that's a risk that you have to take into account, not adjust out of your stats. If I have to make any kind of adjustment, I'd rather just ignore the game completely. Usually I'll do this if a player in Week 16 or 17 is only on the field for one drive or one quarter. But anything beyond that is overcomplicating things. Remember the KISS rule -- keep it simple. Don't spend so much time perfecting an analysis that will never be perfect. Don't discard an analysis as imperfect when non-analysis is even more so.

 

[Chase Stuart]

For Parker it's not too bad, but look at weeks 5 and 9. He had only 15 total carries those two weeks; Bettis had 17 carries in week 5, Staley had 15 in week 9.

Parker's opponents those weeks were GB and SD. GB was exactly average last year in FP allowed to RBs, while SD was pretty stingy. So do we credit Parker a whole game for each of those weeks? A half game? How does that work out.

Treat them as one whole game. There's no need to subdivide it any further. Do you adjust your QB stats when a QB throws 15 passes one game and 35 passes the next? The fact that Parker is in a RBBC is already accounted for in his PPG average. I suppose you could adjust based on the number of carries (similar to how FootballOutsiders do their stats), but you're looking at his game performance. If he's up against a tough defense and only has six carries, then so be it. You don't adjust the San Diego's Points Allowed stats just because one team ran 10 times against them and another team ran 25 times. Points Allowed is Points Allowed.

With QBs, we can very accurately state that Carson Palmer played 15.1 games last year, that Chris Simms played 10.4 games, and that Marc Bulger played 6.8. With RBs it is a bit dicier. I think it gets even worse with WRs.

I don't like to do fractional games, if at all possible. If he played, it's a whole game. If a player misses a half due to injury, that's a risk that you have to take into account, not adjust out of your stats. If I have to make any kind of adjustment, I'd rather just ignore the game completely. Usually I'll do this if a player in Week 16 or 17 is only on the field for one drive or one quarter. But anything beyond that is overcomplicating things.

Remember the KISS rule -- keep it simple. Don't spend so much time perfecting an analysis that will never be perfect. Don't discard an analysis as imperfect when non-analysis is even more so.

It depends what you're ultimate goal is. My goal with the QB study was to determine how much better thant he league average QB each QB was. To do that, partial games must be considered. Injury risk is irrelevant to this analysis -- I want to study only two things: what the guy did on the field and who he did it against.

 

[AB in DC]

The way he did it was right. It can get confusing when you're talking about points allowed. But if the 16 teams on Team X's schedule allowed 15 PPG to RBs, and the league average allowed to RBs is 16.5, then RB Y faced a schedule where the average team RB would score 15 PPG, not 16.5 PPG (the average team RB by definition must score 16.5 PPG). Therefore you would want to increase his total by 110%

Yes.

(or 1.5 FPs, which I'd prefer I think).

No, you can't do this without allocating the 1.5 FPs/G among all of the Pittsburgh's RBs. That gets very complicated when you've got four different RBs who all played a different set of games. That's why it's easier to just multiply by the ratio.

 

[Bri]

Basically, yes, but for all positions.  I don't know why these kind of numbers wouldn't be just as informative as for RB, WR, and TE, even without the kind of analysis you did last year.

It's a bit complicated but the main reason is there's only one QB on any given play. The nature of football makes analyzing RBs, WRs and TEs a bit more difficult when it comes to rear view SOS. I know a few other people feel the same way, so feel free to give us any suggestions on how you think the numbers should be run for those positions. I'm definitely open to suggestions and there's a decent chance I'll have an article on this at some point in the summer.

Just use percentages.Suppose that Pittsburg's opponents averaged 15.0 PPG to RBs, vs. a leaguewide average of 16.5 PPG. (I'm making those numbers up.) In looking at Willie Parker's performance, I'd want to multiply his actual PPG by 16.5/15.0, or 110%, to get an adjusted PPG for 2005.

For those of us who use 2005 stats as the baseline for 2006 projections (as I do), I'd want to do this for every RB to see which players weren't really as good as their numbers would indicate, and which players were a lot better.

In this hypothetical case, Wilile Parker was really 10% better than his numbers due to a tough schedule.

I think your math is wrong and it should be lower. 1.5 less than avg not 110% or more than average

The way he did it was right. It can get confusing when you're talking about points allowed. But if the 16 teams on Team X's schedule allowed 15 PPG to RBs, and the league average allowed to RBs is 16.5, then RB Y faced a schedule where the average team RB would score 15 PPG, not 16.5 PPG (the average team RB by definition must score 16.5 PPG). Therefore you would want to increase his total by 110% (or 1.5 FPs, which I'd prefer I think).

I gotcha, thanks for explaining

 

[Clayton Gray]

What I'd really like to see is a retroactive SOS analysis. It's very hard to predict opponents' performance for the upcoming season, but it's easy to look back at the previous season, see who faced easier or harder defenses vs. their position, and adjust accordingly.

That's a cool idea. I hope I can take a look at that very soon.

 

[AB in DC]

That's a cool idea. I hope I can take a look at that very soon.