[Dynasty] Quality Years Remaining (QYR) (1 Viewer)

[FUBAR]

Player--Team--Draft--Rnd/Pk--Age--QYR

Leinart, Matt ARI Q-06 R1/10 26.7 7.5

you lost me here, Leinart has 7.5 quality years remaining? He probably doesn't have 1.

You could say a lot of players have no quality years remaining depending on how you define quality, but I'm making no attempt to assess a player's value here, only his longevity. Also, as I said in the original post, QYR is only a general guideline and not an individual evaluation of any specific player. I'm simply applying a formula based on averages to all players in a class. Its purpose is to give a rule of thumb number to each class of player (i.e., QB, RB, etc.) based on average peak and end points for that class. My original study of player ages used only players that met a certain standard of fantasy-worthiness (because that's all I really cared about at the time), and end point ages are going to be somewhat higher for such players than for lesser players who flame out earlier. Admittedly, if I apply the formula to all players that someone might have rostered (as I have done in the above post), it ends up overstating QYR for the lesser players. By how much I don't know. That's a flaw, and it's why I have wanted to revisit QYR for lesser players to come up with different end points and multipliers for these category B and C guys since they are generally out of the league earlier than category A guys would be. For now though, the one formula for everyone is what I've got.

The multiplier makes sense. I don't know if it makes more sense to use a projection of points each year or simply tier the player and give points per tier - Chris Johnson at 25 years old, tier 1; give 100 points for tier 1 and 6 quality years remaining, for 600 points. A player like Marion Barber has 4 QYR and projects as a tier 4 RB; if we deduct a standard 15 points per tier, that's 55 points X 4 QYR = 220 points. Now, Michael Turner looks to have 3 QYR and projects as a tier 2 RB. 85 X 3 = 255. We get what we probably already know, Turner is worth more than Barber, but it isn't by much. The biggest difference in various people's rankings will be where they project a player to remain for his QYR and how much they deduct per tier drop. Interesting though, right now we don't have many RBs with only a few years left in the top tier. Perhaps the right way to do this is look at players we'd likely consider equal at different ends of their career? For example, SJax is the oldest guy I have in tier 1 with 4 years left (400 pts). He's probably worth about the same as Chris Wells (arguably) If Wells is a tier 3 with 8 years left and we want to make this equal 400, it would have to be 50 points for tier 3 - 25 points off per tier? I think this gets us in the ballpark.

Your idea is interesting. I will want to look at it more closely when there's more time. Some parts of the valuation and NPV thing that I want to do are still pretty fuzzy in my mind. I have some idea what I want to do, but I'm a guy who has to plow through something and get really immersed into it before I begin to understand it very well. I'm not real intuitive. I wish I were a whole lot smarter than I am. Man, if I had Drinen's IQ.... Oh well. Gotta run, get my day going. Seeya!

My concept might be too simple as it doesn't account for movement between tiers - CJ is certainly a tier 1 for the next 3 years, but in year 5 is he? Conversely, Wells might be tier 3 for now and tier 1 or 2 in a year or two... this could get overly complicated if we let it and IMO it's best to keep it simple as the tier we see them at for the next 3 years.

 

[lyon812]

Massive thanks to CP for his early morning update.

I agree with Fubar about the tiers, though--particularly with so many fluid RB situations at the moment. Projected production and longevity are two disparate variables, since projections are really a combination of talent + opportunity. We tend to base projections on the former, but the latter can change very quickly.

 

[Frank Costanza's Lawyer]

:X

 

[gheemony]

Couch Potato,

Sorry to hear about your father. You have your priorities straight. Wishes to you and your family for good health and good fortune in the coming years.

In your absence, I came up with an NPV formula based on the premise that current year is worth 50%, next season is worth 33%, and 2 seasons from now is worth 17%. I don't think anything past three years is worth much because of uncertainty.

So rough formula would look like this:

PFP10 = FP projections for 2010

PFP11 = PFP10*QYR11/QYR10 - you assume a percentage increase or decrease accordingly

PFP12 = PFP11*QYR12/QYR11 (which is still based on PFP10)

NPV = .50*PFP10 + .33*PFP11 + .17*PFP12

You can effectively have a fixed NPV number for each position and each age, just like a QYR number.

How is this useful? Examples:

1. you have two players projected for the same points in 2010. Who do you take first in a keeper/dynasty league? Look at the NPV. You want the one with the higher NPV.

2. I have an older player projected for more points in 2010 than a younger player. Who's worth more today. Take a look at their NPV. This will show you how much you can expect the older player to regress and younger player to progress.

It was eye-opening to me to see when my instinct on these points didn't match up with the NPV math.

If you want to test this out, I recommend trying to compare the NPV for Peyton Manning, Tom Brady, and Drew Brees with that for Big Ben, Rodgers, Rivers, and Ryan. At what point does age make up for differen in immediate production?

 

[lyon812]

Leinart, Matt ARI Q-06 R1/10 26.7 7.5Warner, Kurt ARI Q-.94 UDFA 38.6 -2.1...Norwood, Jerious ATL R-06 R3/79 26.5 4.3Snelling, Jason ATL R-07 R7/244 26.1 4.6Turner, Michael ATL R-04 R5/154 27.9 3.2

Been playing around with this fascinating methodology, and thought I'd point out that the numbers graciously shared by CP seem to use older formula instead of the newest, unless I have it plugged into Excel wrong.For example, Leinart: 29-(0.828*26.7) = 6.9Warner: 29-(0.828*38.6) = -3.0Norwood: 25-(0.760*26.5) = 4.9Snelling: 25-(0.760*26.1) = 5.2Turner: 25-(0.760*27.9) = 3.8etc.Another thing I noticed is that the ages are a bit off. They seem to correspond to their age in decimals as of 1/1/10, not 9/1/09, which might influence things a bit. Example: Leinart listed age 26.7; birthdate 5/11/83. So he's 26 and ~4 months on 9/1/09, or 26.3.Delhomme listed age 35; birthdate 1/10/75. So he's 34 and ~8 months on 9/1/09, or 34.7.Garrard listed age 31.9; birthdate 2/14/1978. So he's 31 and ~7 months on 9/1/09, or 31.6.I mention this not to criticize CP at all, but because there seem to be a few people that are working on tweaking the formula that might not have noticed. I'm guessing CP's spreadsheet constantly updates the ages, and that they're all current as of today (rather than 9/1).

 

[Couch Potato]

Leinart, Matt ARI Q-06 R1/10 26.7 7.5Warner, Kurt ARI Q-.94 UDFA 38.6 -2.1...Norwood, Jerious ATL R-06 R3/79 26.5 4.3Snelling, Jason ATL R-07 R7/244 26.1 4.6Turner, Michael ATL R-04 R5/154 27.9 3.2

Been playing around with this fascinating methodology, and thought I'd point out that the numbers graciously shared by CP seem to use older formula instead of the newest, unless I have it plugged into Excel wrong.For example, Leinart: 29-(0.828*26.7) = 6.9Warner: 29-(0.828*38.6) = -3.0Norwood: 25-(0.760*26.5) = 4.9Snelling: 25-(0.760*26.1) = 5.2Turner: 25-(0.760*27.9) = 3.8etc.Another thing I noticed is that the ages are a bit off. They seem to correspond to their age in decimals as of 1/1/10, not 9/1/09, which might influence things a bit. Example: Leinart listed age 26.7; birthdate 5/11/83. So he's 26 and ~4 months on 9/1/09, or 26.3.Delhomme listed age 35; birthdate 1/10/75. So he's 34 and ~8 months on 9/1/09, or 34.7.Garrard listed age 31.9; birthdate 2/14/1978. So he's 31 and ~7 months on 9/1/09, or 31.6.I mention this not to criticize CP at all, but because there seem to be a few people that are working on tweaking the formula that might not have noticed. I'm guessing CP's spreadsheet constantly updates the ages, and that they're all current as of today (rather than 9/1).

The first difference you found is an oops on my part. I was tinkering with end points earlier for reasons that aren't important to explain, and forgot to put them back where they belonged before posting above. You are right that Leinart s/b 6.9, Norwood 4.9, etc. Concerning the 2nd difference, yes my SS keeps ages up-to-date automatically, so the calc ends up being as of today, not 9/1/09 or 9/1/10.So, what I'll do is edit post #41 above to correct the oops on the end points, and also to use ages as of 9/1/10 since what you guys would want to see is QYR as of the start of next season, not as of today. Leinart for example becomes QYR of 6.4 based on the future date rather than 6.9 as of today. By the way, if you're setting this up at home on a SS, to get ages exactly right remember to account for leap years, so the age calc is (9/1/10 minus birth date) / 365.25This shouldn't take long. Thanks for catching that.

 

[smashingsilver]

Couch Potato,Sorry to hear about your father. You have your priorities straight. Wishes to you and your family for good health and good fortune in the coming years.In your absence, I came up with an NPV formula based on the premise that current year is worth 50%, next season is worth 33%, and 2 seasons from now is worth 17%. I don't think anything past three years is worth much because of uncertainty.So rough formula would look like thisFP10 = FP projections for 2010PFP11 = PFP10*QYR11/QYR10 - you assume a percentage increase or decrease accordinglyPFP12 = PFP11*QYR12/QYR11 (which is still based on PFP10)NPV = .50*PFP10 + .33*PFP11 + .17*PFP12You can effectively have a fixed NPV number for each position and each age, just like a QYR number. How is this useful? Examples:1. you have two players projected for the same points in 2010. Who do you take first in a keeper/dynasty league? Look at the NPV. You want the one with the higher NPV.2. I have an older player projected for more points in 2010 than a younger player. Who's worth more today. Take a look at their NPV. This will show you how much you can expect the older player to regress and younger player to progress.It was eye-opening to me to see when my instinct on these points didn't match up with the NPV math.If you want to test this out, I recommend trying to compare the NPV for Peyton Manning, Tom Brady, and Drew Brees with that for Big Ben, Rodgers, Rivers, and Ryan. At what point does age make up for differen in immediate production?

I can't get these calculations to work properly. The basic formula for NPV makes sense (NPV = .50*PFP10 + .33*PFP11 + .17*PFP12), but it's the formulas to obtain PFP11 and PFP12 that appear to be flawed, at least to me.PFP11 = PFP10*QYR11/QYR10: The end result of this number will always be less than PFP10, because QYR11 will always be less than QYR10, creating a multiplier of less than 1. I'm not sure if your "assume a percentage increase or decrease accordingly" means that you have to manually add a different multiplier for each player, but if it does, then IMO that's too much work. Maybe I'm just reading this incorrectly, but if you could clarify that statement I would appreciate it. Otherwise, you'll always have players who will score less in future years than in current years, even if their career performance trend is moving upwards. Since your NPV formula already puts less value on the 2011 and 2012 projections, I don't see the point in devaluing future years in the PFP calculations as well.Perhaps a better option would be to compare a player's current age against the peak age for that position, and adding in a multiplier depending on which side of the peak age a player is on. For example, for QBs the peak age listed by CP is 29. If a player is younger than 29, then each succeeding year his PFP should go up (theoretically, of course, since this entire discussion relies on generalizations). If a player is older than 29, then each succeeding year his PFP should go down. I think there are some problems with that, but at least it should provide a more accurate basis for future year projections. Ideally, I think a peak age range should be used, since most players tend to level out at peak performance for a period of years (longer for QBs, shorter for RBs). For QBs, maybe a range of 28-32 could be considered the peak range, where stats could be expected to be consistent over that period of time. Your formula would now need to use an IF statement, but it would probably provide a more accurate projection for NPV. Also, your formula only works if a player's QYR is greater than zero. Once a player's value gets to zero the math doesn't work (can't divide by zero), or if it's a negative number, future years' projections actually increase, not decrease. I'm not sure how to get around that one.

 

[FantasyTrader]

Also, your formula only works if a player's QYR is greater than zero. Once a player's value gets to zero the math doesn't work (can't divide by zero), or if it's a negative number, future years' projections actually increase, not decrease. I'm not sure how to get around that one.

Yeah, but at that point, you don't really need QYR. You already know the player is living on borrowed time, particularly dynasty formats. For practical purposes, once a player reaches 0 QYR, you should analyze the state of your roster without him in the equation - even if he's your #1 QB on any given Sunday that he's on the field.

 

[gheemony]

Couch Potato,Sorry to hear about your father. You have your priorities straight. Wishes to you and your family for good health and good fortune in the coming years.In your absence, I came up with an NPV formula based on the premise that current year is worth 50%, next season is worth 33%, and 2 seasons from now is worth 17%. I don't think anything past three years is worth much because of uncertainty.So rough formula would look like thisFP10 = FP projections for 2010PFP11 = PFP10*QYR11/QYR10 - you assume a percentage increase or decrease accordinglyPFP12 = PFP11*QYR12/QYR11 (which is still based on PFP10)NPV = .50*PFP10 + .33*PFP11 + .17*PFP12You can effectively have a fixed NPV number for each position and each age, just like a QYR number. How is this useful? Examples:1. you have two players projected for the same points in 2010. Who do you take first in a keeper/dynasty league? Look at the NPV. You want the one with the higher NPV.2. I have an older player projected for more points in 2010 than a younger player. Who's worth more today. Take a look at their NPV. This will show you how much you can expect the older player to regress and younger player to progress.It was eye-opening to me to see when my instinct on these points didn't match up with the NPV math.If you want to test this out, I recommend trying to compare the NPV for Peyton Manning, Tom Brady, and Drew Brees with that for Big Ben, Rodgers, Rivers, and Ryan. At what point does age make up for differen in immediate production?

I can't get these calculations to work properly. The basic formula for NPV makes sense (NPV = .50*PFP10 + .33*PFP11 + .17*PFP12), but it's the formulas to obtain PFP11 and PFP12 that appear to be flawed, at least to me.PFP11 = PFP10*QYR11/QYR10: The end result of this number will always be less than PFP10, because QYR11 will always be less than QYR10, creating a multiplier of less than 1. I'm not sure if your "assume a percentage increase or decrease accordingly" means that you have to manually add a different multiplier for each player, but if it does, then IMO that's too much work. Maybe I'm just reading this incorrectly, but if you could clarify that statement I would appreciate it. Otherwise, you'll always have players who will score less in future years than in current years, even if their career performance trend is moving upwards. Since your NPV formula already puts less value on the 2011 and 2012 projections, I don't see the point in devaluing future years in the PFP calculations as well.Perhaps a better option would be to compare a player's current age against the peak age for that position, and adding in a multiplier depending on which side of the peak age a player is on. For example, for QBs the peak age listed by CP is 29. If a player is younger than 29, then each succeeding year his PFP should go up (theoretically, of course, since this entire discussion relies on generalizations). If a player is older than 29, then each succeeding year his PFP should go down. I think there are some problems with that, but at least it should provide a more accurate basis for future year projections. Ideally, I think a peak age range should be used, since most players tend to level out at peak performance for a period of years (longer for QBs, shorter for RBs). For QBs, maybe a range of 28-32 could be considered the peak range, where stats could be expected to be consistent over that period of time. Your formula would now need to use an IF statement, but it would probably provide a more accurate projection for NPV. Also, your formula only works if a player's QYR is greater than zero. Once a player's value gets to zero the math doesn't work (can't divide by zero), or if it's a negative number, future years' projections actually increase, not decrease. I'm not sure how to get around that one.

You're absolutely right. Sorry, I just went off the top of my head.And you're right about the solution. I don't have my spreadsheet, so I'm going off the top of my head again ... My recollection is that I converted QYR into a "percent of peak" value. 100 at the peak age, and some number smaller than that for each age on each other side. I believe that CP's formula takes into account a decline age. I also think that a google search for Doug Drinen and Football Outsiders would give you info about peak and decline ages too. I put these into Excel and used the "best fit" feature to generate a quadratic formula to determine the percent of peak. It's obviously not perfect because it assumes improvement towards peak at the same rate of decline, but it's the best we have. I don't recall seeing any research on approach to peak rates.I'm pretty sure I posted these formulas somewhere in the shark pool. i think if you searched all posts under my ID, you could find them.When i finally break out my spreadsheet, I'll try to remember to post the formulas I use. May not be until Spring. This year, I am trying to do subjective rankings first, and then add in qualitative measures. I find that I let numbers impact my thinking too much, so I am going to hide them for my first set of rankings.

 

[smashingsilver]

Couch Potato,Sorry to hear about your father. You have your priorities straight. Wishes to you and your family for good health and good fortune in the coming years.In your absence, I came up with an NPV formula based on the premise that current year is worth 50%, next season is worth 33%, and 2 seasons from now is worth 17%. I don't think anything past three years is worth much because of uncertainty.So rough formula would look like thisFP10 = FP projections for 2010PFP11 = PFP10*QYR11/QYR10 - you assume a percentage increase or decrease accordinglyPFP12 = PFP11*QYR12/QYR11 (which is still based on PFP10)NPV = .50*PFP10 + .33*PFP11 + .17*PFP12You can effectively have a fixed NPV number for each position and each age, just like a QYR number. How is this useful? Examples:1. you have two players projected for the same points in 2010. Who do you take first in a keeper/dynasty league? Look at the NPV. You want the one with the higher NPV.2. I have an older player projected for more points in 2010 than a younger player. Who's worth more today. Take a look at their NPV. This will show you how much you can expect the older player to regress and younger player to progress.It was eye-opening to me to see when my instinct on these points didn't match up with the NPV math.If you want to test this out, I recommend trying to compare the NPV for Peyton Manning, Tom Brady, and Drew Brees with that for Big Ben, Rodgers, Rivers, and Ryan. At what point does age make up for differen in immediate production?

I can't get these calculations to work properly. The basic formula for NPV makes sense (NPV = .50*PFP10 + .33*PFP11 + .17*PFP12), but it's the formulas to obtain PFP11 and PFP12 that appear to be flawed, at least to me.PFP11 = PFP10*QYR11/QYR10: The end result of this number will always be less than PFP10, because QYR11 will always be less than QYR10, creating a multiplier of less than 1. I'm not sure if your "assume a percentage increase or decrease accordingly" means that you have to manually add a different multiplier for each player, but if it does, then IMO that's too much work. Maybe I'm just reading this incorrectly, but if you could clarify that statement I would appreciate it. Otherwise, you'll always have players who will score less in future years than in current years, even if their career performance trend is moving upwards. Since your NPV formula already puts less value on the 2011 and 2012 projections, I don't see the point in devaluing future years in the PFP calculations as well.Perhaps a better option would be to compare a player's current age against the peak age for that position, and adding in a multiplier depending on which side of the peak age a player is on. For example, for QBs the peak age listed by CP is 29. If a player is younger than 29, then each succeeding year his PFP should go up (theoretically, of course, since this entire discussion relies on generalizations). If a player is older than 29, then each succeeding year his PFP should go down. I think there are some problems with that, but at least it should provide a more accurate basis for future year projections. Ideally, I think a peak age range should be used, since most players tend to level out at peak performance for a period of years (longer for QBs, shorter for RBs). For QBs, maybe a range of 28-32 could be considered the peak range, where stats could be expected to be consistent over that period of time. Your formula would now need to use an IF statement, but it would probably provide a more accurate projection for NPV. Also, your formula only works if a player's QYR is greater than zero. Once a player's value gets to zero the math doesn't work (can't divide by zero), or if it's a negative number, future years' projections actually increase, not decrease. I'm not sure how to get around that one.

You're absolutely right. Sorry, I just went off the top of my head.And you're right about the solution. I don't have my spreadsheet, so I'm going off the top of my head again ... My recollection is that I converted QYR into a "percent of peak" value. 100 at the peak age, and some number smaller than that for each age on each other side. I believe that CP's formula takes into account a decline age. I also think that a google search for Doug Drinen and Football Outsiders would give you info about peak and decline ages too. I put these into Excel and used the "best fit" feature to generate a quadratic formula to determine the percent of peak. It's obviously not perfect because it assumes improvement towards peak at the same rate of decline, but it's the best we have. I don't recall seeing any research on approach to peak rates.I'm pretty sure I posted these formulas somewhere in the shark pool. i think if you searched all posts under my ID, you could find them.When i finally break out my spreadsheet, I'll try to remember to post the formulas I use. May not be until Spring. This year, I am trying to do subjective rankings first, and then add in qualitative measures. I find that I let numbers impact my thinking too much, so I am going to hide them for my first set of rankings.

Good advice. I'll do some extra research and play around with the formulas.

 

[Danish Bengal]

Good threads cannot be bumped too often

 

[Biabreakable]

Wow, my old thread got bumped again. I'd better chime in here and not be so lazy.

 

[trader jake]

With the changes in the NFL rulebook over the past few seasons, it may be time to re-analyze the quarterback remaining years index. The old men at that position appear to be making the original number a bit obsolete with their 35+ age production.