With five weeks on the books, we're just starting to see the first wave of owners look at their record and their team and decide to throw in the towel on the season. And inevitably the first question they ask is “if I'm rebuilding, should I hold on to _______ or sell him?”
My answer is always exactly the same, (the same as my answer to every trade question, in fact): sell him if the value of what you're getting back exceeds the value of what you're giving up. Otherwise, don't.
I recognize that this can be unhelpful advice, but I see so many owners dumping quality starters like Jordy Nelson for a pittance just because Nelson is old and they are rebuilding and they think the proper way to rebuild is to sell all your old players for whatever you can get.
With that said, players at different points in their career have different values depending on your team's goals. It'd be disingenuous to suggest that older players don't become less valuable to you the minute you switch into rebuild mode. You ultimately should be willing to accept less for Jordy Nelson after you decide to rebuild than you were willing to accept before you decide to rebuild.
The trick is deciding how much less to accept. Several years ago, I wrote about thinking about player aging like a mortality table, with players not declining so much as abruptly falling off. From that concept, I built actual mortality tables for each position based on historical data from 1985 to 2014. Here they are for quarterbacks, running backs, wide receivers, and tight ends.
And from those tables, I can calculate what percentage of a player's career value is lost as they become another year older.
(One important caveat: there are two decisions that impacted these tables. First, by using data from 1985 to 2014 I was able to build a larger sample, but I probably underrated the impact that modern training and medicine have on longevity, especially at quarterback. Second, to end the tables I used an approach called “capping” where I eventually just set the mortality rate to 100%. The result is that these tables overrate the value lost by older skill players to a small degree and by older quarterbacks to a large degree.)
Caveats properly emptored, let's look at the raw data. Here's every position, as well as what percentage of value remains at each age if you no longer care about this season.
| Age | QB | RB | WR | TE |
|---|---|---|---|---|
| 21 | 0.93 | 0.89 | 0.91 | 0.89 |
| 22 | 0.92 | 0.88 | 0.90 | 0.88 |
| 23 | 0.92 | 0.87 | 0.89 | 0.88 |
| 24 | 0.91 | 0.86 | 0.89 | 0.88 |
| 25 | 0.91 | 0.84 | 0.88 | 0.87 |
| 26 | 0.90 | 0.83 | 0.87 | 0.87 |
| 27 | 0.89 | 0.81 | 0.86 | 0.86 |
| 28 | 0.89 | 0.79 | 0.85 | 0.86 |
| 29 | 0.88 | 0.76 | 0.83 | 0.85 |
| 30 | 0.87 | 0.72 | 0.81 | 0.84 |
| 31 | 0.86 | 0.65 | 0.79 | 0.83 |
| 32 | 0.84 | 0.54 | 0.77 | 0.82 |
| 33 | 0.83 | 0.29 | 0.73 | 0.80 |
| 34 | 0.81 | 0.00 | 0.67 | 0.77 |
| 35 | 0.79 | 0.58 | 0.73 | |
| 36 | 0.76 | 0.38 | 0.68 | |
| 37 | 0.72 | 0.00 | 0.55 | |
| 38 | 0.65 | 0.00 | ||
| 39 | 0.54 | |||
| 40 | 0.30 | |||
| 41 | 0.00 |
So... quod erat demonstrandum, right? Jordy Nelson is 32 years old, a 32-year-old receiver only has 77% as much value if you no longer care about this season, therefore when you switch into rebuild mode Jordy Nelson automagically loses 23% of his value to you. Right?
Well... no. You see, these figures assume you value all seasons equally. Would you trade a 2019 first round draft pick for a 2089 first-round draft pick, though? No, no you would not.
The reason you would not is that we do something called “time discounting”. The further away a season is, the less we value it. This is a normal and rational decision— leagues fold or unforeseen circumstances come up that cause you to leave. There's a good chance you're not still playing fantasy football in 2089, let alone still playing in that same dynasty league.
There's a good discussion to be had about how much we should time-discount, but I've found 20% seems to be a good amount that lines up reasonably well with how owners value future draft picks in real leagues. (If anything, I'd say the industry-wide implied time discount is actually closer to 30%, but I think 20% is probably more appropriate.)
How does a time-discount impact these numbers? Well, consider the example of a 21-year-old quarterback. By going from 21 to 22, the estimated number of years remaining in his career falls from 13.14 to 12.18. But most of those “extra” years won't be realized for more than a decade, by which point they've been time-discounted so much they're essentially valueless.
Without time discounting, switching into rebuilding mode reduces a 21-year-old quarterback's value by 7%. With a 20% time discount, however, that reduction falls to just 2%. Even if you're rebuilding, there's very little practical difference between how you value a 21-year-old quarterback and a 22-year-old quarterback.
So here's one last chart, providing the same information as before except with a 20% time discount applied.
| Age | QB | RB | WR | TE |
|---|---|---|---|---|
| 21 | 0.98 | 0.94 | 0.96 | 0.94 |
| 22 | 0.98 | 0.93 | 0.95 | 0.93 |
| 23 | 0.97 | 0.92 | 0.95 | 0.93 |
| 24 | 0.97 | 0.91 | 0.94 | 0.92 |
| 25 | 0.96 | 0.89 | 0.93 | 0.92 |
| 26 | 0.96 | 0.87 | 0.92 | 0.91 |
| 27 | 0.95 | 0.85 | 0.91 | 0.90 |
| 28 | 0.94 | 0.83 | 0.89 | 0.90 |
| 29 | 0.93 | 0.79 | 0.88 | 0.89 |
| 30 | 0.92 | 0.75 | 0.86 | 0.87 |
| 31 | 0.91 | 0.68 | 0.83 | 0.86 |
| 32 | 0.89 | 0.56 | 0.80 | 0.84 |
| 33 | 0.88 | 0.30 | 0.76 | 0.82 |
| 34 | 0.85 | 0.00 | 0.70 | 0.79 |
| 35 | 0.83 | 0.60 | 0.76 | |
| 36 | 0.79 | 0.39 | 0.70 | |
| 37 | 0.75 | 0.00 | 0.57 | |
| 38 | 0.68 | 0.00 | ||
| 39 | 0.56 | |||
| 40 | 0.30 | |||
| 41 | 0.00 |
With the time discount, Jordy Nelson, (as a 32-year-old receiver), now loses 20% of his value instead of 23%. It's still a substantial drop— you're essentially wiping out a fifth of his dynasty value the second you enter rebuild mode— but I'd wager it's much smaller than most would have expected. (Especially when you keep in mind that anyone you trade him for will have also lost some percentage of their remaining value when you switched into rebuilding mode.)
Which is why I say most owners are too willing to sell low on aging studs when they rebuild. As long as you'll be competitive as early as next year, they're capable of still providing you with plenty of value going forward.