Chris Black (@ByChrisBlack) just penned a very interesting piece built around a $200 million question. Should the Raptors re-sign Kyle Lowry to a long term deal at near-max money? The article takes the approach of comparing the total WS (catch-all win production stat from basketball-reference) earned by guards under 6’4” at ages 31+ versus ages below 30, to suggest a likely post-age-30 level of production relative to the player’s pre-age-30 production.
It’s a good broad look at how guards tend to age, but I think the result (24% of career WS coming after age 30) is misleading, suggesting that the player will only be one third as productive in his post-30’s years as his average pre-age-30 year.
There are really two questions that need to be asked about a player heading into his later seasons. Will he play in all of them? And how good will he be when he does play?
How Long Will Lowry Play?
If the Raptors re-sign Kyle Lowry this coming summer, they can offer up to a five year deal, while other teams can offer up to a four year deal. So, what are the odds he plays for the full four or five years?
So, we’ll start the same way Black did — find all the guards 6’4” or shorter, sort them by most WS produced, and choose a cut-off. But we’ll differ right away here — instead of filtering by most WS produced before age 30, we’ll focus on how guards play right before they hit 30, in their prime. Since we are interested in how Lowry will perform up to potentially his age 35 season, we’ll look at ages 25-30 to capture a player’s peak. In this case, using a cutoff of about 30 total WS generated in a player’s peak gives a list of 64 players, where Lowry ranks 31st in WS. A nice bounding set where Lowry is the “average” player (see here for full BR report).
Then, since that data set was six years of performance, we’ll pull the next six years for each of those players to compare to (similarly sized samples allows for easier comparisons). So, ages 31-36. That also gives a slight buffer on the age 35 target we have for Lowry. Here is the second set of data (much bigger, as I’ve applied no WS cutoff to ensure we have results for everyone in the first set). Some players have limited samples in this set because they are still playing and haven’t hit 36 years old yet. As such, I removed any players actively playing from the sample, leaving us with a sample of 52 guards who generated 30 WS in their prime.
So, to answer the first question: how long is Lowry likely to play? Well, of the 52 guards in the sample, the average number of years played after the age of 31 is 4.35 years. But that is skewed by a few players who stopped at age 30 or 31, which Lowry shows no sign of doing. The most common result for the sample? Playing six full years after the age of 31. Here’s a look at how many guards in our sample played a certain number of years.
6 years or more: 40%
5 years: 12%
4 years: 19%
3 years: 13%
2 years: 6%
1 year: 6%
0 years: 4%
That’s looking pretty good for longevity. Lowry only needs to reach five years (if the Raptors sign him to a 5-year deal), which based on the above, he’d have a 52% chance of doing. Sign him to a 4-year deal instead, and that jumps to 71%.
So that might show that perhaps the Raptors would be better off giving Lowry only a 4-year deal, but it seems to suggest that most guards at least keep playing for 4-5 years of their 31+ age seasons.
But Will He Play Well?
Playing games is one thing, actually producing entirely another. So, using the same sample chosen above, how much WS generation did over-31 players maintain relative to their peak years?
Looking at the 25-30 sample and the 31-36 sample, and finding the total WS generated per year for each player, we can generate a WS retention value, where WS retention is the average yearly WS in the age 31-36 sample divided by the average yearly WS in the age 25-30 sample.
For example, John Stockton had an average WS of 13.7 per year in his prime, and 11.4 from age 31 to 36. So his WS retention number would be 0.83. So if Lowry replicates Stockton’s success, he would be expected to generate win shares at 83% the rate he has in his prime. Obviously not everyone is like Stockton though. So let’s look at how the data broke down by breaking it down into blocks — how many players had retention rates of 0%, 50%, 100%?
>100% WS Retention: 8%
90%-100% WS Retention: 8%
75%-90% WS Retention: 2%
50%-75% WS Retention: 23%
25%-50% WS Retention: 42%
10%-25% WS Retention: 8%
0%-10% WS Retention: 10%
There are a few outliers at the top and bottom end, but the safe bet appears to be right in the middle, between 25% and 75% WS retention. The average WS retention value of the entire group is almost exactly 50% anyway (49.898% is pretty close) so we’ll use that as our baseline expectation.
So, over Lowry’s age 25-29 seasons (as this year is barely started), he averaged 8.2 WS per season. Dividing that by two gives an expected WS production of 4.1 WS per season in the future. Now, this is pretty conservative, as he has this year to add to that age 25-30 sample, and he hasn’t dipped below 7 WS in any of the past three seasons. If we assume that this season he puts up an average of the past three seasons since he became a full time starter, that’s a 10 WS season this year, and puts his expected WS in his age 31-36 seasons at 4.3 WS per season.
But what is that worth? Well, the simplest argument is that there are 30 teams in the league, and 82 games each, meaning there are 41 wins for each team on average. So take the total salaries, divide that by 30 and by 41, and you get the average cost of a “win.” One WS should be valued the same as a win, and so you can come to an estimate of how much a player will be worth.
The current CBA (who knows what will happen with the new one) assumes that the average team salary will be roughly 12% above the cap. The projected cap levels are shown below.
2017-18 cap: $103 million
2018-19 cap: $108 million
2019-20 cap: $109 million
2020-21 cap: $114 million
2021-22 cap: $118 million (estimate, no projection has been provided by the league)
That means an average salary cap of $110 million over the course of Lowry’s deal. Which means an average team salary of $123.4 million. And that results in the value of a single win (or WS) being almost exactly $3 million.
Generally teams pay more for their older established players than they are worth, so being over or under their contract value isn’t a big concern, but ideally you’d like to be in the ball park.
Unfortunately for Lowry (or the Raptors, depending how you look at it), his 4.3 projected WS per season gets valued at a $13 million salary. Which is pretty far below his maximum salary (which would average out at $39 million over a 5-year deal, or $37.5 million over a 4-year deal).
If the Raptors sign Lowry to a deal like that, they are betting on him replicating his top notch play through his age 34-35 season. It’s not the typical result for a short guard, but it’s not impossible or even impractical. As shown above, 16% of the sample, or about 1 in 6 guards, maintained 90% or more of their peak performance, and of those guards, they averaged a 106% retention rate (several guards actually got better with age, Sam Cassell, Steve Nash, Mark Jackson among them). Applying that 106% retention rate to Lowry’s peak average performance puts him at 9.1 WS per season, which is worth about $27 million. That’s about $10 million below his projected maximum salary, but such is life in a league with rookie scale contracts skewing salary-to-performance ratios. That’s at least in the ball park.
For the record, if you use Lowry’s last few years (age 27-29) as his peak rather than from age 25, as he’s been a late bloomer, that projected performance goes up to 10.7 WS per year, worth about a $32 million salary. That’s a very generous assumption though.
All of these numbers projected are far above the expectation of generating WS at only one third the production rate of his career to date. But the most likely case is not better enough, at roughly three quarters of his career average to date, to be worth re-signing at maximum money.
If the Raptors believe Lowry will age very well, and can replicate the performance retention of the top one sixth of the guards in the sample, then he is worth paying to keep around, although it’s a far safer bet to give him four years than five. But if they don’t trust in his ability to hold up at a far above average rate, then they can’t expect him to produce nearly enough to be worth his salary, and will have to think long and hard about giving him a lot of money.