The 2021 OPS Projections Thread

[Frobby]

Bringing back a tradition I've skipped the last few years.    I've collected the Orioles players' OPS projections from various projection systems and put them on the spreadsheet below.    Then I made my own assumptions about how the plate appearances will be allocated, using an assumption of 6200 PA total, and used those to come up with a team projected OPS.   As you'll see, the projected team OPS ranges from a low of .705 (THE BAT) to a high of .726 (ZiPS).    

A few of these projection systems are new to me: ATC, THE BAT, and THE BAT X.   They are all published on Fangraphs, along with ZiPS and Steamer.   Marcel is published on BB-Ref, and PECOTA is only available to BP members but I was able to get it from an anonymous source (he knows who he is!).   I may try to add a few others later.

If anyone wants to post his/her own OPS projections for these 18 players, I'll be happy to plug them in and let everyone know what team OPS results, based on my PA assumptions.    There are several players who I expect will get more PA than I listed if they are healthy all year, but I've hedged a little for injury risk.  If anyone thinks I'm wildly off in how I allocated the playing time, please feel free to comment!

Name 		PA	PECOTA	ZiPS	Steamer	ATC	THE BAT	THEBATX	Marcel
Severino	350	0.704	0.694	0.708	0.701	0.703	0.705	0.699
Sisco		200	0.714	0.725	0.696	0.717	0.711	0.713	0.707
Rutschman	150	0.603	0.664	0.674	0.691	0.705	0.705	0.674
Mancini		550	0.793	0.823	0.791	0.801	0.798	0.813	0.806
Davis		150	0.632	0.565	0.603	0.594	0.611	0.608	0.576
Sanchez 	425	0.661	0.694	0.667	0.683	0.645	0.672	0.672
Valaika		325	0.706	0.783	0.677	0.681	0.658	0.663	0.686
Jones		100	0.639	0.597	0.645	0.642	0.646	0.645	0.758
Ruiz		500	0.704	0.704	0.727	0.701	0.685	0.672	0.706
Bannon		175	0.673	0.681	0.694	0.693	0.692	0.691	0.687
Galvis		600	0.713	0.682	0.672	0.692	0.684	0.677	0.708
Mountcastle	550	0.741	0.798	0.788	0.782	0.759	0.755	0.802
Hays 		450	0.755	0.747	0.741	0.731	0.713	0.711	0.771
Mullins 	300	0.637	0.686	0.689	0.666	0.651	0.659	0.681
Santander 	550	0.769	0.796	0.784	0.804	0.781	0.791	0.769
Stewart 	175	0.719	0.748	0.764	0.762	0.729	0.735	0.753
Diaz		150	0.698	0.731	0.721	0.716	0.701	0.699	0.711
Others		500	0.649	0.649	0.649	0.649	0.649	0.649	0.649
TOTAL	   	6200	0.709	0.726	0.716	0.718	0.705	0.708	0.722

 

[Frobby]

Just for a little context here, the Orioles' team OPS was .725 in 2019 and .750 last year.   So, these systems are projecting a step back.

[ChosenOne21]

Wow, Marcel loves Jahmai Jones and ZIPS loves Valaika

These seem low on Mountcastle, and i have a feeling if Rutschman is playing well enough to make the majors he'll hit better than those projections. I think there's a pretty good chance Hays outperforms these projections as well. 

Rest seems fairly reasonable

[ChipTait]

So, the newest iteration of the “M&M Boys” - Mancini & Mountcastle - should be hitting third and fourth most of the year, I would guess.

I think your projected number of PAs for Chris Davis may be “optimistic” at 150.  I think we will see more of him than 1/4th of a season worth of plate appearances.

[survivedc]

If Stewart does end up hitting to a .730-50 OPS I would be surprised to seem him limited to 175 plate appearances.

Out of curiosity, do any of these services have best case/top end type projection? 

[wildcard]

The thing that I think is true here is the O's will have 4 guys at the bottom of the order that can not hit well.   Catchers, Ruiz, Galvis and Sanchez.   These guys are probably rally killers. Last year they had a surprise from Iglesias and Alberto who killed lefties.   

Better defense, worse offense from these 4 positions combined

[ChosenOne21]

11 hours ago, Frobby said:

Just for a little context here, the Orioles' team OPS was .725 in 2019 and .750 last year.   So, these systems are projecting a step back.

Don't these systems always tend to be conservative by nature? Predicting players to be worse than they end up being more likely than not?

[Frobby]

8 hours ago, ChosenOne21 said:

Don't these systems always tend to be conservative by nature? Predicting players to be worse than they end up being more likely than not?

I’d have to look into that.   One thing I think is true is that however you think the PA’s will get allocated, they end up getting adjusted according to performance.    So if Mullins has a .637 OPS as PECOTA forecasts, he probably doesn’t end up with 300 PA.    

Offhand, I do think ZiPS tends to be offense-friendly, but I’ll have to investigate.  
 

[Just Regular]

9 hours ago, ChosenOne21 said:

Don't these systems always tend to be conservative by nature? Predicting players to be worse than they end up being more likely than not?

Those that give one line it is usually a 50th percentile line,   I know when PECOTA was new to me, that was the first time I saw one publish its 90/80/70/60/50/40/30/20/10.

I know Clay Davenport recently incorporated NCAA stuff into his forecasts (not sure if totally finished), but he also shows each 10% on his Player Cards and Adley is now up to a ~300/400/500 line on his 90th percentile (I think that percentile just gives rates and everyone 162-games, so wouldn't say the 8-win part of it).

http://www.claydavenport.com/ht/RUTSCHMAN19980206A.shtml

[Frobby]

I can only find three years of past threads, though I’m pretty sure I did more.   In any event:

2015: Actual .728 (.730 league average), ZiPS .731, Steamer .725, PECOTA .708, Average .725

2016: Actual .760 (.744), ZiPS .746, Steamer .737, PECOTA .744, Average .741

2017: Actual .747 (.753), ZiPS .750, Steamer .756, Average .748

I put in the league average since these projection systems really can’t predict how the offensive context of the league is going to change over time.  The projection systems correctly projected that the offense would be better in 2016 than in 2015, but really we improved more than projected, even adjusting for the change in league average.   In 2017 they correctly projected a downturn.    

[Frobby]

1 hour ago, OrioleDog said:

Those that give one line it is usually a 50th percentile line,   I know when PECOTA was new to me, that was the first time I saw one publish its 90/80/70/60/50/40/30/20/10.

I know Clay Davenport recently incorporated NCAA stuff into his forecasts (not sure if totally finished), but he also shows each 10% on his Player Cards and Adley is now up to a ~300/400/500 line on his 90th percentile (I think that percentile just gives rates and everyone 162-games, so wouldn't say the 8-win part of it).

http://www.claydavenport.com/ht/RUTSCHMAN19980206A.shtml

The more interesting part to me is he has Rutschman at .737 OPS with 25 HR and 92 RBI on his 10th Percentile projection line.   Honestly if he’s that good, he should be in the majors on April 16 when he no longer qualifies for a full year of service.    

[ChosenOne21]

Hmmm, those systems were closer to reality than I remember them being.

I think what I meant to say is that the systems have a hard time accounting for player breakthroughs. If a guy has a couple years of like .730 OPS, then something clicks and he puts up an .880, the projection systems are much more likely to put his projected OPS closer to the .730 than the .880, even if the breakthrough is real and he's likely to put up OPS in the high 800s going forward. Basically, the system is more likely to write off the performance as a fluke than being representative of the player's true talent. Sure, players put up fluky seasons all the time, but they also have breakthroughs.

Look at Santander. The projection systems believe last year was a fluke. Maybe they're right, I guess we'll see

[Frobby]

46 minutes ago, ChosenOne21 said:

Hmmm, those systems were closer to reality than I remember them being.

I think what I meant to say is that the systems have a hard time accounting for player breakthroughs. If a guy has a couple years of like .730 OPS, then something clicks and he puts up an .880, the projection systems are much more likely to put his projected OPS closer to the .730 than the .880, even if the breakthrough is real and he's likely to put up OPS in the high 800s going forward. Basically, the system is more likely to write off the performance as a fluke than being representative of the player's true talent. Sure, players put up fluky seasons all the time, but they also have breakthroughs.

Look at Santander. The projection systems believe last year was a fluke. Maybe they're right, I guess we'll see

The projections think Santander played 37 games last year.   That’s not a lot.   

[Frobby]

Updated to (1) add Maikel Franco, (2) substitute Urias for Sanchez, and (3) add in my own projections (which, needless to say, are more optimistic than any of the others!).    If any one else would like to make OPS projections for each of the listed players, I will plug them into the spreadsheet later and let you know what team OPS emerges.

Name 		PA	PECOTA	ZiPS	Steamer	ATC	THEBAT	THEBATX	Marcel	Frobby
Severino	350	0.704	0.694	0.708	0.701	0.703	0.705	0.699	0.721
Sisco		200	0.714	0.725	0.696	0.717	0.711	0.713	0.707	0.749
Rutschman	150	0.603	0.664	0.674	0.691	0.705	0.705	0.674	0.721
Mancini		550	0.793	0.823	0.791	0.801	0.798	0.813	0.806	0.821
Davis		100	0.632	0.565	0.603	0.594	0.611	0.608	0.576	0.575
Urias		300	0.708	0.761	0.729	0.692	0.675	0.675	0.768	0.701
Valaika		300	0.706	0.783	0.677	0.681	0.658	0.663	0.686	0.721
Jones		100	0.639	0.597	0.645	0.642	0.646	0.645	0.758	0.681
Franco		450	0.756	0.724	0.782	0.753	0.722	0.725	0.743	0.745
Ruiz		225	0.704	0.704	0.727	0.701	0.685	0.672	0.706	0.695
Bannon		125	0.673	0.681	0.694	0.693	0.692	0.691	0.687	0.681
Galvis		600	0.713	0.682	0.672	0.692	0.684	0.677	0.708	0.701
Mountcastle	550	0.741	0.798	0.788	0.782	0.759	0.755	0.802	0.785
Hays 		450	0.755	0.747	0.741	0.731	0.713	0.711	0.771	0.825
Mullins 	300	0.637	0.686	0.689	0.666	0.651	0.659	0.681	0.681
Santander 	550	0.769	0.796	0.784	0.804	0.781	0.791	0.769	0.801
Stewart 	175	0.719	0.748	0.764	0.762	0.729	0.735	0.753	0.745
Diaz		150	0.698	0.731	0.721	0.716	0.701	0.699	0.711	0.705
Others		475	0.649	0.649	0.649	0.649	0.649	0.649	0.649	0.649
TOTAL	   	6100	0.718	0.732	0.726	0.724	0.711	0.713	0.732	0.738

 

[BohKnowsBmore]

1 hour ago, Frobby said:

Updated to (1) add Maikel Franco, (2) substitute Urias for Sanchez, and (3) add in my own projections (which, needless to say, are more optimistic than any of the others!).    If any one else would like to make OPS projections for each of the listed players, I will plug them into the spreadsheet later and let you know what team OPS emerges.

Name 		PA	PECOTA	ZiPS	Steamer	ATC	THEBAT	THEBATX	Marcel	Frobby
Severino	350	0.704	0.694	0.708	0.701	0.703	0.705	0.699	0.721
Sisco		200	0.714	0.725	0.696	0.717	0.711	0.713	0.707	0.749
Rutschman	150	0.603	0.664	0.674	0.691	0.705	0.705	0.674	0.721
Mancini		550	0.793	0.823	0.791	0.801	0.798	0.813	0.806	0.821
Davis		100	0.632	0.565	0.603	0.594	0.611	0.608	0.576	0.575
Urias		300	0.708	0.761	0.729	0.692	0.675	0.675	0.768	0.701
Valaika		300	0.706	0.783	0.677	0.681	0.658	0.663	0.686	0.721
Jones		100	0.639	0.597	0.645	0.642	0.646	0.645	0.758	0.681
Franco		450	0.756	0.724	0.782	0.753	0.722	0.725	0.743	0.745
Ruiz		225	0.704	0.704	0.727	0.701	0.685	0.672	0.706	0.695
Bannon		125	0.673	0.681	0.694	0.693	0.692	0.691	0.687	0.681
Galvis		600	0.713	0.682	0.672	0.692	0.684	0.677	0.708	0.701
Mountcastle	550	0.741	0.798	0.788	0.782	0.759	0.755	0.802	0.785
Hays 		450	0.755	0.747	0.741	0.731	0.713	0.711	0.771	0.825
Mullins 	300	0.637	0.686	0.689	0.666	0.651	0.659	0.681	0.681
Santander 	550	0.769	0.796	0.784	0.804	0.781	0.791	0.769	0.801
Stewart 	175	0.719	0.748	0.764	0.762	0.729	0.735	0.753	0.745
Diaz		150	0.698	0.731	0.721	0.716	0.701	0.699	0.711	0.705
Others		475	0.649	0.649	0.649	0.649	0.649	0.649	0.649	0.649
TOTAL	   	6100	0.718	0.732	0.726	0.724	0.711	0.713	0.732	0.738

 

Any chance you can add a column with a simple average of all of the projection systems and sort on that?