The importance of baseball projections cannot be understated when it comes to your fantasy drafts. In a recent article, I analyzed the accuracy of various projection sources for fantasy purposes. I had found that my Special Blend of projections fared the best in 2018. But, as I typically do, I still got to work on trying to tweak them further based off my research. The 2019 projections should hopefully be another step forward for this aggregate projection system.
My Special Blend projections blend data from a variety of free projection sources on the market as well as calculated xStats and stats from previous years. Based off of years of research, each individual statistic is given different weighting and combinations. When I first started doing fantasy baseball research, I noticed that just simply combining multiple projections produced more accurate results. As I grew older and wiser, I realized that certain projection systems were better than others at specific stats. This projection system grew out of that idea. Each year, I do analysis using linear regression to find which combination of projections would have done best in the previous year. That helps me continue to mold and alter the weights of my Special Blend projections and produce more accurate aggregated projections.
One of the main adjustments in the 2019 version of my projections is that I have removed the Fangraphs Fan projections from my blending. A major weakness in those particular projections is that they had a very limited player pool and it prevented me from projecting the entire fantasy baseball draft universe. By removing the Fan projections, it has added hundreds of new players to my mix. In addition, I have continued to add xStats and other sources to the blend. The full list of sources in this aggregate are:
- Steamer
- CAIRO
- ZiPS
- Clay Davenport
- Fangraphs Depth Charts
- THE BAT
- 2018 xStats
- 2018 Actual Stats
If those sources did not exist, my Special Blend would not exist so many thanks to them for all of their hard work in generating their initial projections.
The 2019 Projections
I’ve embedded the 2019 projections below, but you can download them for yourself by accessing them through the Google Doc here. There are separate tabs for pitcher and hitter projections. The projections also include a calculated WERTH roto value for each player based on a standard 5×5 roto league format. The WERTH value is a z-score that calculates how far above or below average a player is in the main roto stats. If you want custom WERTH values for your specific league settings, check out my Excel cheatsheet where you can input your league settings.
One Final Note
These will be added into my cheatsheets very soon. I will be doing an update to the cheatsheets in the near future to address various bug fixes, as well as getting the latest data in there. I will announce on Twitter when they are available.
Projections last updated on 03/27/2019
Effem
03/11/2019 at 9:34 AMThanks! How do you estimate playing time?
Ken
03/11/2019 at 9:45 AMThanks Luke! One thing though, it appears Edwin Diaz is missing from your pitcher projections. Unless you foresee a precipitous drop off with him on the Mets this year.
Luke Gloeckner
03/11/2019 at 11:19 AMAh, that’s interesting. He is missing a projection from Clay’s projections so wasn’t in the Special Blend. I’m sure he has one though so I’ll see why that didn’t carry over and get him added in. Thanks!
Ken
03/11/2019 at 11:49 AMThanks, also, if you still are looking for them, I have ATC projections from Feb, 2018. Hitters and pitchers.
Jason
03/11/2019 at 12:07 PMHey Luke. Great work. Like Ken had mentioned, as I drop this into my Excel sheet and try to vlookup it over to my main tab, there are several players missing. So far I’ve come across Vlad Jr., Edwin Diaz, Yusei Kikuchi, Andrew Miller, Craig Kimbrel, Will Smith, Carlos Martinez, Dallas Keuchel, Adam Jones. I see some of the trends there (free agents, injuries, etc), but just wanted to point out some things you may have missed. Like I said though, great work. Definitely using these over anything else!
Luke Gloeckner
03/11/2019 at 1:47 PMGood catches! I’ll look into those a bit more and see if I can pinpoint why they’re missing a projection.
Luke Gloeckner
03/12/2019 at 7:27 PMThanks again for catching these. Was able to correct most of those (a couple of them are still missing projections in at least one of the systems).
Curious George
03/11/2019 at 12:58 PMDrafting this Friday, think the updated cheatsheets will be up by then?
Luke Gloeckner
03/13/2019 at 4:13 PMYup, I just uploaded version 1.60 today, which includes these projections.
AH
03/11/2019 at 6:24 PMTiny suggestion: You can freeze the top row of the Google Doc (select it, View>Freeze>1 row) so the categories follow you down the page. Makes reading stat lines easier.
Luke Gloeckner
03/12/2019 at 7:28 PMWeird, yeah. Forgot to set that. Thanks for the reminder. Also, love the avatar. Texas forever.
Bobby Mueller
03/12/2019 at 1:32 AMLuke, thanks for putting these projections together. I’m concerned about your projections for on-base percentage, though. Almost every player seems low, by roughly 3-5%. I just grabbed a bunch of players and averaged their OBP projections from Steamer, ZiPS, FanGraphs Depth Charts, THE BAT, and Clay Davenport. I don’t have CAIRO or xStats included, but when I compare your projected OBPS with the average of the five I listed, yours are always lower. Here are a few examples:
Francisco Cervelli:
.330–Mr. Cheatsheet
.357–Steamer
.357–ZiPS
.357–FGDC
.356–THE BAT
.347–Clay
Justin Turner
.352–Mr. Cheatsheet
.370–Steamer
.376–ZiPS
.373–FGDC
.379–THE BAT
.368–Clay
Randal Grichuk
.286–Mr. Cheatsheet
.298–Steamer
.300–ZiPS
.299–FGDC
.296–THE BAT
.306–Clay
Khris Davis
.312–Mr. Cheatsheet
.323–Steamer
.326–ZiPS
.324–FGDC
.332–THE BAT
.341–Clay
Anyway, perhaps you could check on this and make sure you haven’t done something wrong somewhere along the way. Thanks again.
Luke Gloeckner
03/12/2019 at 10:04 AMCool, I’ll check it out. Could be that the OBP calculation was missing a parameter. The baseline numbers (walks, PAs, etc) should all be fine.
Luke Gloeckner
03/12/2019 at 10:25 AMOkay, yeah, I see the mix-up. I won’t be able to upload the updated results until tonight but thanks for pointing it out!
Luke Gloeckner
03/12/2019 at 7:29 PMFYI, got it fixed and uploaded with the correct OBP values. Thanks again.
Bobby Mueller
03/13/2019 at 2:35 PMThanks for doing that. Glad I could help.
Erick
03/12/2019 at 12:00 PMHey Luke,
Thanks for the great work man!
when are you going to have the sleeper hitters available?
Thanks again!
Tom
03/13/2019 at 4:23 PMGreat work. I feel like a complete idiot asking this but when I click on the google link how do I actually download the rankings? It just opens a web page.
Luke Gloeckner
03/13/2019 at 4:30 PMHaha, that’s a completely reasonable question. I didn’t realize I had it linked to the table view. I updated the link. Try clicking again and it should take you to a sheet where you can download it.
Joe Thomas
03/13/2019 at 5:10 PMOur draft is March 23. How many updates do you think you will do between now and then.
P.S. Love what you have done so far. Glad the Google pointed me to your site.
mike
03/14/2019 at 7:12 AMHi Luke,
Thank you for all of your efforts in putting this together. I am confused by your projection on judge, if he plays the whole year, I don’t see how he only hits 35 hrs… if you compare him to stanton, they play in the same ballpark, k about the same, about the same fly ball rate and Judges average fb is actually longer than stantons and distance on hr’s is about the same. He hit 52 the year before, understand he had an insane hr/fb ratio , but even if you regress to 30. still hits 40.
Also he scuffled when returning from broken wrist last year, which drove his stats down a little. Says he is now fully healthy which seems to be the case with 4hrs in only 26 abs this spring.
Luke Gloeckner
03/14/2019 at 11:13 AMWelp, my projection system blends together the work of other projection systems and xStats. They’re not my own personal opinions. Both Steamer and ZiPS project Judge for 35 HRs over 619 and 560 PA respectively, for instance. Is he capable of outperforming his projections? Of course.
Sal
03/14/2019 at 4:24 PMHi Luke – Any idea when the tiers cheatsheet might be available
JB
03/15/2019 at 1:15 PMHi Luke – Great job with the spreadsheets.
I have another suggestion for you to consider.
Many of the websites have different position eligibility. Can you make that an option for your spreadsheets to take into account?
Ex: we can choose yahoo! or ESPN or CBS for position eligibility.
Then maybe your spreadsheet can incorporate a column on the ‘draft board’ next to ‘notes’ that shows the players eligibility for the league we are playing in…
It’s very helpful to know if we can put a player at another position while we are drafting.
Thanks and keep up the great work!
Joe
03/18/2019 at 5:39 PMFirst time using the CheatSheet. I’m really blown away by how nice this is. I love playing around with it . Its so versatile. I’m sure this represents an enormous amount of work and finesse with Excel. One question. Do you have as much in the Pts. League spreadsheet as the Rotos
Brian
03/21/2019 at 9:12 PMSuggestion: Add ages to the special blend projections? They are in the Cheatsheet, but I prefer to make my own colorizing notations from the data here- and the Excel Cheatsheet is locked (which is probably good, since it would be easy to bork the scripts).
Thanks!
hankp
03/23/2019 at 7:55 PMLuke
Do you update what is available on the projections tab when you update the special blend? Thanks!