Hot Button Hitters: The Divisive Projections
A couple of days ago, I released some hitter rankings and now I’m going to dive into some players with the biggest value divergence. By going through the most different players, the value averaging makes any effects of an outlier less meaningful. Also, remember that these are the most divisive projections. Other player projections involve less disagreement.
It’s not always one projection being “wrong”. Each one as some instance of diverging. By using several projections, the faults of one are lessened.
Now, I could easily be wrong in the following assessments. Which of the following projection do you think is wrong based on 600 PA? Please let me know in the comments.
Note: Again, I’m not giving out who are the other sources as the focus is just looking for differences. Also, all the projections or pro-rated to 600 plate appearances to make sure the differences are in talent, not playing time.
Charlie Blackmon (25th in NFBC ADP)
| Source | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| FG Depth Chart | 90 | 23 | 70 | 11 | .287 |
| Source 1 | 104 | 25 | 72 | 12 | .303 |
| Source 2 | 102 | 27 | 69 | 12 | .301 |
| Source 3 | 96 | 26 | 73 | 9 | .282 |
| Max | 104 | 27 | 73 | 12 | .303 |
| Average | 98 | 25 | 71 | 11 | .293 |
| Min | 90 | 23 | 69 | 9 | .282 |
Our Depth Charts and Source 3 hate Blackmon mainly with Runs and AVG. I can interpret where all but the Runs values come from, though. Here are his past three seasons all prorated to 600 PA:
| Year | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| 2016 | 104 | 27 | 77 | 16 | .324 |
| 2017 | 113 | 31 | 86 | 12 | .331 |
| 2018 | 103 | 25 | 60 | 10 | .291 |
He’s going to be 32-years-old so some aging factor numbers will be dragging him down, thus a drop is expected. All the projections, except the .282 AVG, seem fine but once it gets average in, it isn’t out of place.
I’m just more of 100 Runs guy. I wondered if the team context changed much, but it hasn’t. In 2018, the Rockies average 4.79 Runs/Game and are projected for 4.84 Runs this year. The average projection lessens the effect of the low Depth Chart value, so I’m fine moving forward with his averaged value.
Nelson Cruz (129th in NFBC ADP)
| Source | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| FG Depth Chart | 87 | 35 | 101 | 1 | .282 |
| Source 1 | 80 | 33 | 91 | 1 | .260 |
| Source 2 | 79 | 35 | 100 | 1 | .274 |
| Source 3 | 83 | 36 | 110 | 1 | .278 |
| Max | 87 | 36 | 110 | 1 | .282 |
| Average | 82 | 35 | 100 | 1 | .274 |
| Min | 79 | 33 | 91 | 1 | .260 |
Source 1 is not a fan of Cruz and just doesn’t expect the 100 RBI or the batting average. Cruz did post an .256 AVG last season fueled by a .264 BABIP but normally has an AVG near .280 with a BABIP over .300. I’m perfectly fine going with my average value.
Khris Davis (45th in NFBC ADP)
| Source | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| FG Depth Chart | 82 | 36 | 98 | 1 | .240 |
| Source 1 | 92 | 43 | 112 | 1 | .248 |
| Source 2 | 80 | 31 | 96 | 2 | .255 |
| Source 3 | 89 | 42 | 109 | 2 | .246 |
| Max | 92 | 43 | 112 | 2 | .255 |
| Average | 86 | 38 | 104 | 1 | .247 |
| Min | 80 | 31 | 96 | 1 | .240 |
These projections are all over the place with the Depth Charts and Source 2 being on the low pessimistic side. The .255 projected AVG from Source 2 seems to be the worst offender. Before diving in, here are his stats from the past three seasons.
| Year | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| 2016 | 84 | 41 | 100 | 1 | .247 |
| 2017 | 84 | 40 | 101 | 4 | .247 |
| 2018 | 90 | 44 | 113 | 0 | .247 |
It seems like that the two higher projections are weighing 2018 more while the other two are taking 2016 and 2017 into account more. As a 31-year-old, a production drop should be expected so the median value seems about right especially since it puts his AVG at .247.
Miguel Andujar (68th in NFBC ADP)
| Source | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| FG Depth Chart | 76 | 25 | 87 | 4 | .279 |
| Source 1 | 71 | 23 | 70 | 3 | .265 |
| Source 2 | 81 | 26 | 91 | 4 | .293 |
| Source 3 | 74 | 23 | 88 | 3 | .289 |
| Max | 81 | 26 | 91 | 4 | .293 |
| Average | 76 | 24 | 84 | 4 | .282 |
| Min | 71 | 23 | 70 | 3 | .265 |
What a cluster f___ with the batting average. It’s almost a 30-point range with Source 2 being Mr. Optimist and Source 1 hating his guts. The .265 AVG is just completely out of place and by going with the average it’s not as bad.
Tommy Pham (64th in NFBC ADP)
| Source | Runs | HR | RBI | SB | AVG |
|---|---|---|---|---|---|
| FG Depth Chart | 77 | 21 | 69 | 17 | .256 |
| Source 1 | 104 | 25 | 72 | 20 | .270 |
| Source 2 | 95 | 19 | 64 | 16 | .268 |
| Source 3 | 100 | 25 | 73 | 20 | .273 |
| Max | 104 | 25 | 73 | 20 | .273 |
| Average | 94 | 22 | 70 | 18 | .267 |
| Min | 77 | 19 | 64 | 16 | .256 |
Our depth charts have some issues with Pham and it’s mainly with the program doesn’t have him leading off and getting all the runs scored. I’m also, not sure what’s up with the batting average hate but some may be linked to a 26 K%. Again, the averaged value seems appropriate but I’ll need to re-examine the Run and RBI mix.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
Be honest, you only wrote this article because even Khris Davis’s average averages hit .247 and you needed an excuse to bring it to light. I thank you for doing so.
LOL. I missed that – that’s funny. Both the the stat itself and the fact that you got to write “average averages”
Both Davis (.247 BA) and Nomar Mazara (20 HR) are sitting on 3 year streaks of the exact same number in a category and have the once-in-a-lifetime opportunity to join Adam Dunn (40 HR from 2005-2008) as players to have posted the exact same number in a stat 4 years in a row. I’m sure media coverage will be extensive. . . .
EDIT**** I had it in my head that Davis was at 3 years, but it looks like he’s actually at 4 years in a row, though he didn’t qualify for the batting title in 2015. Not sure how the “rule book” will treat that. Does that get an asterisk? 🙂
Davis has a 4 year streak of .247 avg, started with brewers in 2015.
Was 2018 not Khris Davis’s fourth year of .247?
busted
Something that stands out about Blackmon is his wOBA has outpaced his xwOBA by a considerable amount each of the past four seasons. In fact, 2018 represented his smallest gap, yet he still out OBA’d his expected wOBA by .27. In the prior two seasons, his 2016 wOBA (.394 > .357) and his 2017 (.415 > .377) were in the top 3% and 2% of MLB, respectively. I don’t look at these numbers enough to know whether this is a Coors-influenced pattern you’d expect to see across pretty much any competent Rockies hitter.
Blackmon may be deceptively old. His birthday is July 1st. He would be entering his age 33 season if he were born one day earlier.
It is odd seeing his runs scored projected to decline so much, as he led the NL in the category in each of the last two seasons. Are projection systems plugging in a considerably worse hitter than LeMahieu to bat behind him next season?
Finally, Blackmon’s K rate has gradually ticked up closer to league average. He’s run a BABIP well above .300 in each of his six full seasons in MLB. In the three seasons it exceeded .350, he batted above .300, sometimes well above .300. In the three seasons it fell closer to .300, his batting average dipped to approx. .290 (.288, .287 and 2018’s .291). His two best seasons for BABIP/AVG were in 2016-2017, age 30-31. These two seasons look an awful lot like his peak, and the aging curve supports that.
I think Blackmon’s a fine player, and Source 3 may be a little quick to cover themselves in case he doesn’t age gracefully. I feel like the 2016-2017 AVG may be overweighted if you use three years of data for the projections. I’d take FG on that one since a younger Blackmon has been a 50/50 bet to land at .290. If any of the systems saw his peak period as almost totally over, the average AVG would fall below it.
xwOBA really needs to add both a ballpark component and a speed component. Blackmon benefits from both.
Of the 8 Rockies with at least 250 PA last year, 6 of them exceeded their xwOBA by an average of 33 points. The 2 who didn’t (DJLM and Ianetta) were barely below their xwOBA by .003 and .016.
Blackmon’s averages per 600 PA the last three years:
Runs: 107
OBP: .380
HR: 27.6
Steamer600 projections:
Runs: 90
OBP: .354
HR: 23
I think the reduction in runs can be reasonably explained by the corresponding reduction in OBP and home runs in the projection, plus the runs data might be somewhat noisy and prone to regression in general.
I originally thought this article would explore outlier projections, how often the outlier is correct, and discuss the dangers of averaging projections. You went another way with it and that’s cool, but I’d love to read that article I had in mind.
Do any of the sources factor in batting order position in the projections? If so, that could lead to divergence. I ask because the Fan projections here uses batting order and from playing around with different slots I can tell those projections are definitely materially impacted by using different batting slots.