Saric rookie comparisons

[Crow]

It might be an issue with how the macro ran on my computer / Excel version or something. If you didn't change something, it is probably just a use error.

[JoshEngleman]

This is so fascinating that I had to share it. I decided to take a look at Ben McLemore. Unsurprisingly, things looks bleak for Ben. The one silver lining on his entire list is C.J. McCollum. Buddy Hield showing up 6th is hilarious. I wonder much differently we would look at Ben if played 4 years at Kansas. He would be a rookie, just like Buddy.

[JoshEngleman]

I decided to finally make some changes that I had been meaning to try. Instead of just looking at a particular season, I'm applying the same weighting principles to each stat. For example, instead of just looking at Isaiah Thomas's 2016-17 TS%, I'm going with a a weighted TS% over the past 3 years. No idea if this is good or bad, but the results are certainly slightly different. It looks a bit more accurate from a positional standpoint, but I don't really know if there's a right answer or not. Here are a few guys using this new method.

Isaiah Thomas



Al Horford


Dario Saric


J.J. Redick

[Crow]

KCP wants $20 plus mil. Might get it. How you weight things matters on "worth". One year strong 3 point fg% might justify the salary request. 3 yrs of data and you probably say no. RPM has him around plus 1. Pay $20 mil for that? Some do but it is not a big value score at that level. May be bad deal, depending who else you can get and at what price.

[JoshEngleman]

Here's what I have now for KCP. I don't think he'll have any problem getting $20M/per next year, and that opinion has nothing to do with these sim scores. Someone is going to give him a big offer sheet.

[Crow]

Where / how did you get S4 mil. for a current win? By just comparing average team payroll to 41 avg. wins I get $2.5 - 3 mil. per win. $4 mil. might be ok for wins over replacement level players or maybe I am forgetting something.

I probably wouldn't give KCP $20 mil. I don't see a case for him helping make a team better than average unless maybe he is an expensive 3rd wheel. He hasn't made it happen in Detroit. If you don't have / can't get a true #1 or #2, maybe you fake it with him but I wouldn't expect that good results.

[JoshEngleman]

Combo of a few things. I had tried to dive into calculating the $/win a while back to deal with the big cap spike, so I used that as a baseline. I originally had it as $3.5M, but I was seeing higher numbers everywhere else. 538's Carmelo projections are using ~$5.1M, which felt way too high. I was comfortable using $4M right now, but I need to dig into the data a little bit to be fully confident.

I could be totally wrong with my thought process, but the cap spike makes this a bit quirky. When the cap growth is linear, then you can do a quick estimation like you did and it will work reasonably well. The problem here is that 75% of the contracts are on the old scale. People are paying a way higher amount per win because of the non-linear jump of salary. Basically, because of the jump, the literal average $ per win isn't the same as what is being spent in the market.

If I'm the Pistons, I wouldn't want to pay him $20+ million per year. I'd let a team like the Nets make a big offer and I wouldn't match. It's one of those weird situations where I'd retain him at that price if my team was good, but since the Pistons aren't good and I don't expect them to be good, then it's easier to just let him walk.

[JoshEngleman]

Just did a quick sense check. I went to Spotrac and grabbed all of the FA contracts signed before this season. I used the AAV just to make things a bit quicker. 129 guys signed for a total of $1.12 billion. I pro-rated my numbers through a full season for simplicity sake. The 129 guys are expected to produce a total of 308.6 wins. Just using this info, I get to $3.6M per win from last year's spending spree. I think $4M per win looks pretty spot on for '17-18.

[Crow]

Thanks for the response and extra work. But now I feel even stronger about the exchange rate between performance & salary being too high, especially for this season.

I can see using the average value in the free agent market more on future years than the current one but probably always at a shifting blend of the pre-existing contracts vs. the new. in the current season pre-existing deals far outnumber those signed last summer (confirmed by your numbers just provided). For season after this one it is about even, again suggesting a blended figure not total use of the higher one. Going into free agent market is a choice and usually the worse choice for player acquisition, by some over trades on average and by a lot compared to draft or the minimum end of free agent market, or even just the new entrant part of that.

[JoshEngleman]

It all depends what you are trying to measure. When the season is over, it's really easy to see how much teams literally spent on this season's wins, but if we're trying to figure out contracts in the future, then we want to know what team's are actively spending on the open market. I think that $4M number is perfect for right now. It's a great time to be a free agent.

[Crow]

Well, we disagree; but ok. I'll just chop your estimates by about 25% in current year for my consideration / use and maybe by 10% in year 2.

[JoshEngleman]

I think we're actually on the same page. The $4M is more like a predictive number. That's around what we would expect teams to spend this off-season. The true $ spent on wins is likely around $3M per.

[Crow]

Same page about what the numbers say but less so on how to use them. But, again, thanks for the reveal and discussion. I had some hope others with similarity systems might chime in but I mostly should give up on that hope or trying to prod. There is some past threads on the topic you could try to find. From Ed Kupfer, David Sparks, Mike G, probably Lane, I think Statman, maybe K Pelton, maybe someone else I am not remembering.

[jgoldstein34]

Finally got around to running my sim for Dario Saric, a handful of decent comps (Paul George, Luol Deng) in their, but mostly all role players or worse.



The wins metric there is my own per minute wins added stats with replacement level being .426 for some context.

As for the discussion on what a good $/win would be, I used ~$3.04 M for 2016-17. I used this formula to get there:

% per win = ( [ (Salary Cap)/31 )/[ (Salary Cap) ]

I assume a team could get 10 wins just by paying replacement level players, so the $/win is the cost for those 31 wins for an average team (average team wins 41, 41-10=31). The above formula gives me 3.23% of the cap equates to about one win above replacement. Over the next few seasons that gives the follow $/win:

Code: Select all

Year	   Cost per Win
------    -------------
2016-17	$3,036,870.97
2017-18	$3,290,322.58
2018-19	$3,322,580.65
2019-20	$3,488,709.68
2020-21	$3,663,161.29
2021-22	$3,846,322.58

To finish getting an estimated salary, I just multiple the players expected wins by the $/win and then add the minimum salary to adjust for the fact that I use wins above replacement. I'm not sure how accurate this is, but it makes logical sense to me and generally approximates contracts players get in free agency.

[Crow]

Thanks for the input.

How would you say your Wins Added metric varies most significantly from other major media metrics? Which one is it most like?

Do you use weights on your comp system inputs? If so, what gets the most weight? Do use height as a category? Do you use an aging curve? Your own or a published one?