Average Value Theory (1 Viewer)

[BuckeyeArt]

I'm sure this may have been discussed but I haven't seen it. I've seen using AVT for drafting purposes but not for trading. It seems to make sense to use when trading across different positions. For example, what should you be able to get for TE1 or TE2? The AVT should answer this.

So, I looked at historical values and used 5 years (no significant difference from 10). I used a starting team of 1 QB, 2 RB, 3 WR, and 1 TE to determine baselines. I chose a baseline of about 1/3 more than last starter (QB17, RB33, WR49, TE17). The scoring system used was standard with 1/2 PPR (1 for 10 yds R&R, 1 for 20 passing, 6 for R&R TD, 4 for passing, -1 int). From there, I calculated the values over the past 5 years and found the average value for each slot.

So, according to AVT chart below, the TE1 or TE2 should be worth about the same as RB9-11, WR8-11, or QB3. Well, anyone who has ever tried to trade a TE would agree that is not the case. So why is that? I thought maybe it was because people felt the other positions (esp. RB) were more likely to retain that same level of production.

So, the next thing I did was look at the future production (following 3 years) of each player. Values for any year that fell below the baseline were assigned a '0' rather than a negative value (someone not producing or being used as a starter won't hurt you, but they also won't help you). I took the average of those 3 years and then took 5 year averages of each slot. Those are listed in the table below next to the "Average Values" and are titled "Future Values". At the bottom of each column, I listed the correlations between the future values and average values for each position.

The results were interesting. TE's, WR's, and RB's were roughly equal with TE's leading the way. QB's were noticeably lower. That suggests that the performance of TE's is more consistent than that of RB's or WR's (although only slight) and is much more consistent than QB's. So, why is a TE so hard to trade? Of course, the reverse is also true. It should be very easy to acquire one of the top TE's and relatively cheap to do so (compared to that of a top 10 RB or WR, or a top 2 QB) and you can be fairly confident the production will continue.

Any thoughts?

Average Values				 Future Values										  Rnk	QB	 RB	 WR	 TE		QB	 RB	 WR	TE	  										   1	157	244	182	116		52	100	104	76												   2	132	234	151	 98		68	162	 92	56												   3	 97	196	134	 76		61	109	 76	33												   4	 90	181	130	 61		57	103	100	 7												   5	 80	162	124	 54		35	 49	 77	24												   6	 76	150	121	 53		37	133	 54	 3	 										   7	 74	130	117	 44		32	 84	 69	27												   8	 70	116	112	 38		26	 64	 50	11												   9	 65	108	109	 29		43	 39	 78	10												   10	57	102	 99	 25		13	 46	 44	14	   										   11	55	 92	 94	 20		36	 54	 22	 1												   12	50	 87	 92	 17		20	 91	 49	19	  										   13	39	 81	 87	 15		32	 12	 47	 1												   14	33	 80	 86	 12		26	 40	 63	 6												   15	17	 72	 84	  8		65	 28	 34	 8												   16	 9	 68	 78	  4		10	 54	 39	 3											   17	 0	 61	 73	  0		 8	 33	 49	 0	   										   18		   58	 72					  45	 22				 										   19		   53	 68					  27	 35				  										   20		   50	 65					  33	 58				  										   21		   46	 62					  30	 54				  										   22		   40	 61					  12	 50				  										   23		   36	 57					   7	 20				 										   24		   31	 54					  46	  6															   25		   26	 52					   7	 36				   										   26		   23	 50					  47	 55															   27		   20	 48					   7	 35				 										   28		   18	 45					  22	 20				   										   29		   14	 42					   5	 16				   										   30		   11	 39					   9	 12				   										   31			9	 37					  11	 45				   										   32			3	 34					  53	 25														   33			0	 33					   4	  0			  										   34				  30							 21															   35				  29							 32															   36				  28							 24															   37				  24							  6															   38				  22							 23															   39				  22							  0															   40				  20							  8															   41				  18							 15															   42				  15							 21															   43				  13							 15															   44				  12							 27															   45				   8							 42															   46				   6							 20															   47				   4							 21															   48				   3							 11					   										   49				   0							  6																   																			   QB	 RB	 WR	 TE												  Correlation				 0.54   0.83   0.83   0.86

 

[pollardsvision]

this is a good question that you pose.

for the WR part, i'd want to look at the historical ADP's of TE1/2 vs WR8/9. i'm not sure they'd be that different. i know over the last few years, it took a 3rd round pick to get an antonio gates, yet could still get a top 10 WR in the 3rd round. the numbers would help to see if that subjective value gap exists.

the RB gap is true. no one would trade any TE for a top 10 RB. what explains this?

personally, i'd move the RB baseline. i generally consider an RB baseline of at least 36 and usually go higher than that, even up to 42 or 48. i see 36 as a minimum b/c i view the 3rd RB as a player that i'd want to feel comfortable starting. i go further sometimes b/c a 4th RB is likely to be a necessary contributor also.

my value baseline difference, as well as the real-life value difference, can be attributed to need to have good RBs on the bench b/c RB's are likely to get hurt or bust (strangely, the low bust potential of TE's actually decreases their value b/c of this). 3-4 RBs are a necessity. often, there's only a need to have 1 TE on the roster and find somebody for a bye week.

if the RB baseline was around 42, i think the values would fall in line with what we see in leagues.

 

[BuckeyeArt]

this is a good question that you pose.for the WR part, i'd want to look at the historical ADP's of TE1/2 vs WR8/9. i'm not sure they'd be that different. i know over the last few years, it took a 3rd round pick to get an antonio gates, yet could still get a top 10 WR in the 3rd round. the numbers would help to see if that subjective value gap exists.

It feels that way but it's not really true. According to FBGs ADP consensus, the TE1 goes at 40 and TE2 goes at 48. Between those picks, WR15 and WR16 are chosen. WR8-10 go between picks 24 and 28.RB is even worse with RB21-24 going between the 1st 2 TE's. I'll have to see what it does to RB values when you lower the baseline.

 

[KCC]

The results were interesting. TE's, WR's, and RB's were roughly equal with TE's leading the way. QB's were noticeably lower. That suggests that the performance of TE's is more consistent than that of RB's or WR's (although only slight) and is much more consistent than QB's. So, why is a TE so hard to trade?

I haven't done any analysis on this, but my guess would be that the reason TEs appear more consistent is because of a much narrower range in stats for that position due to a lower ceiling. If you look at high end production for RBs, WRs, and QBs, these numbers are going to much greater than average production at those positions than top TE production would be from average TE production.

 

[BuckeyeArt]

my value baseline difference, as well as the real-life value difference, can be attributed to need to have good RBs on the bench b/c RB's are likely to get hurt or bust (strangely, the low bust potential of TE's actually decreases their value b/c of this). 3-4 RBs are a necessity. often, there's only a need to have 1 TE on the roster and find somebody for a bye week.if the RB baseline was around 42, i think the values would fall in line with what we see in leagues.

Moving the RB baseline from 33 to 43 adds about 25-26 points to the RB values. This moves the equivalent point (to TE1-2) to RB12-16. Since the RB's chosen between TE1-2 are RB21-24, the TEs are still being undervalued.

 

[BuckeyeArt]

The results were interesting. TE's, WR's, and RB's were roughly equal with TE's leading the way. QB's were noticeably lower. That suggests that the performance of TE's is more consistent than that of RB's or WR's (although only slight) and is much more consistent than QB's. So, why is a TE so hard to trade?

I haven't done any analysis on this, but my guess would be that the reason TEs appear more consistent is because of a much narrower range in stats for that position due to a lower ceiling. If you look at high end production for RBs, WRs, and QBs, these numbers are going to much greater than average production at those positions than top TE production would be from average TE production.

The correlations were nearly equal. The size of the number shouldn't matter. Since the top TEs tend to remain the top TEs and the top RBs tend to remain the top RBs, the positions should be comparable. In other words, the amount scored at each slot should be roughly consistent from year to year. So TE1-2 should generally have the same value as RB9-16 (depending upon where you set the baseline).

 

[KCC]

The results were interesting. TE's, WR's, and RB's were roughly equal with TE's leading the way. QB's were noticeably lower. That suggests that the performance of TE's is more consistent than that of RB's or WR's (although only slight) and is much more consistent than QB's. So, why is a TE so hard to trade?

I haven't done any analysis on this, but my guess would be that the reason TEs appear more consistent is because of a much narrower range in stats for that position due to a lower ceiling. If you look at high end production for RBs, WRs, and QBs, these numbers are going to much greater than average production at those positions than top TE production would be from average TE production.

The correlations were nearly equal. The size of the number shouldn't matter. Since the top TEs tend to remain the top TEs and the top RBs tend to remain the top RBs, the positions should be comparable. In other words, the amount scored at each slot should be roughly consistent from year to year. So TE1-2 should generally have the same value as RB9-16 (depending upon where you set the baseline).

My bad.I didn't understand you were using the position rankings for your correlation.

 

[EBF]

My answer from a dynasty perspective:

- People overvalue mediocre RBs. We've seen it in recent years with guys like Willie Parker, Laurence Maroney, and LaMont Jordan. These guys aren't bad players, but they're not difference makers and they're easy for NFL teams to replace. Yet you see guys like this go in the top 3-4 rounds of dynasty leagues every year because people hold the misguided notion that because a player is a RB, that player is valuable. The truth is that only the top 10-15 RBs really carry significant long term value. Beyond that point you might as well draft a different position.

- TEs are common. Most leagues require you to start 2-3 RBs. There are only 32 NFL teams in the NFL. What this means is that in the average league, owners will always feel squeezed at the RB position. The short supply relative to the high demand built into the league structure encourages owners to keep an iron grip over any halfway decent RB. On the flipside, most leagues only start 1 TE. With 32 teams in the NFL, there's a big supply of starting TEs for a small number of teams in your league. The result is that you can almost always get a fringe top 10 TE in a trade or draft for pretty cheap.

So while a guy like Winslow or Witten has a lot of value in a dynasty league, they can be difficult to trade because few teams are willing to pay a premium for an elite player at a position where they can probably get a solid starter for dirt cheap.

 

[pollardsvision]

my value baseline difference, as well as the real-life value difference, can be attributed to need to have good RBs on the bench b/c RB's are likely to get hurt or bust (strangely, the low bust potential of TE's actually decreases their value b/c of this). 3-4 RBs are a necessity. often, there's only a need to have 1 TE on the roster and find somebody for a bye week.if the RB baseline was around 42, i think the values would fall in line with what we see in leagues.

Moving the RB baseline from 33 to 43 adds about 25-26 points to the RB values. This moves the equivalent point (to TE1-2) to RB12-16. Since the RB's chosen between TE1-2 are RB21-24, the TEs are still being undervalued.

hmm. what happens when you move the TE baseline to greater account for the difference in the # of RBs need on a roster (5-8) vs the number of TE's (never more than 2). the other factor influencing the TE baseline is the number of similar TE's after the top few.i think the TE baseline can safely be put at 12, possible down to 10. after the 1st 4-6 TEs are taken, there are a boatload of pretty similar TEs out there. a TE baseline of 10-12 makes intuitive sense, considering the number of fantasy teams that will pay $1 for their starting TE (b/c there's not a huge difference b/w, say TE 7 and TE 12). the actual season ending value may be very different, but the peceived value going into the season isn't too different b/w the TEs after the top 4-6.also, looking at this year alone, i wouldn't expect the top TEs to be that highly valued. the top TE, witten, is a player that many expect to have a major regression. i don't think people are seeing a healthy, dominant antonio gates type of difference maker this year.

 

[BuckeyeArt]

So while a guy like Winslow or Witten has a lot of value in a dynasty league, they can be difficult to trade because few teams are willing to pay a premium for an elite player at a position where they can probably get a solid starter for dirt cheap.

The problem with that thought is the word 'solid'. There is nothing solid about the 10th ranked TE. It is equal (in value) to RB26 or WR37. So, I agree that what you said may be the reason but it is flawed.

 

[EBF]

So while a guy like Winslow or Witten has a lot of value in a dynasty league, they can be difficult to trade because few teams are willing to pay a premium for an elite player at a position where they can probably get a solid starter for dirt cheap.

The problem with that thought is the word 'solid'. There is nothing solid about the 10th ranked TE. It is equal (in value) to RB26 or WR37. So, I agree that what you said may be the reason but it is flawed.

Maybe so, but I think people have more faith in their ability to pluck a good sleeper from the ranks of lesser TEs. Witten wasn't a monster until last season. This year you have several guys like Owen Daniels, Alge Crumpler, and Zach Miller who could conceivably put up top 5 numbers. I think people are more comfortable going into the year with something like RB20/TE10 as opposed to RB30/TE3, even if it's actually a mistake.

 

[pollardsvision]

My answer from a dynasty perspective:

- People overvalue mediocre RBs. We've seen it in recent years with guys like Willie Parker, Laurence Maroney, and LaMont Jordan. These guys aren't bad players, but they're not difference makers and they're easy for NFL teams to replace. Yet you see guys like this go in the top 3-4 rounds of dynasty leagues every year because people hold the misguided notion that because a player is a RB, that player is valuable. The truth is that only the top 10-15 RBs really carry significant long term value. Beyond that point you might as well draft a different position.

- TEs are common. Most leagues require you to start 2-3 RBs. There are only 32 NFL teams in the NFL. What this means is that in the average league, owners will always feel squeezed at the RB position. The short supply relative to the high demand built into the league structure encourages owners to keep an iron grip over any halfway decent RB. On the flipside, most leagues only start 1 TE. With 32 teams in the NFL, there's a big supply of starting TEs for a small number of teams in your league. The result is that you can almost always get a fringe top 10 TE in a trade or draft for pretty cheap.

So while a guy like Winslow or Witten has a lot of value in a dynasty league, they can be difficult to trade because few teams are willing to pay a premium for an elite player at a position where they can probably get a solid starter for dirt cheap.

i agree with your point, but i also think that is part of what creates a necessary increase in value of RB's.it's true, there will only be about 15 real difference making RBs.

the problem is, who will they be? we are all regularly wrong about this, whether it's due to injury or just a straight out bust.

people know that they can't confidently go draft 2 RBs and be "set" at the position. if you knew for a fact that you're top 2 RBs would be difference makers, there wouldn't be a strong need to load up on RB3/4's (aka perceived mediocre players).

since it's likely that one of a team's top 2 RB's will bust on them, the value of 3rd and 4th RBs will increase.

i could see waiting if one doesn't see a huge difference b/w RB's 20 and 35, but the need to get them at some point is still there. no other position has this phenomenon (at least not as strong).

TE's, on the other hand, are very different. i think most people feel quite confident that they could predict the top 6 TEs in some order. in theory, this would make the top TEs more valuable, but b/c it decreases the total number of TE's needed, it also decreases the overall value of TE's in general.

 

[BuckeyeArt]

hmm. what happens when you move the TE baseline to greater account for the difference in the # of RBs need on a roster (5-8) vs the number of TE's (never more than 2). the other factor influencing the TE baseline is the number of similar TE's after the top few.i think the TE baseline can safely be put at 12, possible down to 10. after the 1st 4-6 TEs are taken, there are a boatload of pretty similar TEs out there. a TE baseline of 10-12 makes intuitive sense, considering the number of fantasy teams that will pay $1 for their starting TE (b/c there's not a huge difference b/w, say TE 7 and TE 12). the actual season ending value may be very different, but the peceived value going into the season isn't too different b/w the TEs after the top 4-6.

Since I'm using 12 teams, TE12 (or TE13 in this case since I used last starter +1) would be the lowest baseline. Using that, it only lowers the TE values by 15 and moves the RB range of equivalency to RB15-19, still 5-6 slots off what it should be. In addition, using TE12 as the baseline assumes that the TE talent is exactly distributed evenly, which really is an unrealistic assumption.

also, looking at this year alone, i wouldn't expect the top TEs to be that highly valued. the top TE, witten, is a player that many expect to have a major regression. i don't think people are seeing a healthy, dominant antonio gates type of difference maker this year.

The names shouldn't matter since I'm comparing slots. Also, the numbers are 5 year averages so Witten's big year is muted by the average.

 

[BuckeyeArt]

i agree with your point, but i also think that is part of what creates a necessary increase in value of RB's.it's true, there will only be about 15 real difference making RBs.the problem is, who will they be? we are all regularly wrong about this, whether it's due to injury or just a straight out bust.people know that they can't confidently go draft 2 RBs and be "set" at the position. if you knew for a fact that you're top 2 RBs would be difference makers, there wouldn't be a strong need to load up on RB3/4's (aka perceived mediocre players). since it's likely that one of a team's top 2 RB's will bust on them, the value of 3rd and 4th RBs will increase.i could see waiting if one doesn't see a huge difference b/w RB's 20 and 35, but the need to get them at some point is still there. no other position has this phenomenon (at least not as strong).TE's, on the other hand, are very different. i think most people feel quite confident that they could predict the top 6 TEs in some order. in theory, this would make the top TEs more valuable, but b/c it decreases the total number of TE's needed, it also decreases the overall value of TE's in general.

The general problem with that argument is that the correlations are very nearly equal. So while people believe they can predict TEs better, they really can't. RBs are nearly as predictable as TEs.

 

[pollardsvision]

also tied into all this, especially in a non-PPR league, is the fear of literally running out of serviceable options at the RB position.

at times, it feels like RBs are dropping like flies and teams are scrambling just find anybody that will get them some points at the RB position some weeks.

there's nothing worse than have a RB in the starting lineup that you know, beyond a shadow of a doubt, will not produce 3 points.

everyteam is always a few hits away from facing this awful reality.

this is not true of the TE position. you can always find a waiver option with a shot at 50 yards, which is even more meaningful b/c a 50 yd. TE performance will not be much different than many teams in the league each week.

 

[pollardsvision]

i agree with your point, but i also think that is part of what creates a necessary increase in value of RB's.it's true, there will only be about 15 real difference making RBs.the problem is, who will they be? we are all regularly wrong about this, whether it's due to injury or just a straight out bust.people know that they can't confidently go draft 2 RBs and be "set" at the position. if you knew for a fact that you're top 2 RBs would be difference makers, there wouldn't be a strong need to load up on RB3/4's (aka perceived mediocre players). since it's likely that one of a team's top 2 RB's will bust on them, the value of 3rd and 4th RBs will increase.i could see waiting if one doesn't see a huge difference b/w RB's 20 and 35, but the need to get them at some point is still there. no other position has this phenomenon (at least not as strong).TE's, on the other hand, are very different. i think most people feel quite confident that they could predict the top 6 TEs in some order. in theory, this would make the top TEs more valuable, but b/c it decreases the total number of TE's needed, it also decreases the overall value of TE's in general.

The general problem with that argument is that the correlations are very nearly equal. So while people believe they can predict TEs better, they really can't. RBs are nearly as predictable as TEs.

i'm not saying that the value difference is right or wrong. just offering some explanations. it seems like this difference really must just come down to the need for so many RBs.

 

[BuckeyeArt]

i'm not saying that the value difference is right or wrong. just offering some explanations.

I know. I'm just pointing out they're flawed.

it seems like this difference really must just come down to the need for so many RBs.

perceived need.The other interesting thing is that it seems people have generally come to the right conclusion when drafting QB's. They generally wait until late after the first few QBs come off the board. That thinking appears to be correct, given the low correlation of the QB slots to future value (especially the lower ones). People also tend to draft TEs the same way, even though the TE slots are much better correlated with future success.