Introducing THT Forecasts
Just in time for the start of Spring Training, I’m excited to present to you the newest Hardball Times offering—THT Forecasts.
In a partnership with Brian Cartwright, the brilliant inventor of the Oliver projection system, we have produced forecasts for over 7,000 major and minor league players. For each player, Brian has provided us with detailed projections for the next six years, along with raw statistics and major league equivalencies for each of the past three seasons. This includes hitting, pitching, and fielding statistics, and coming soon, base running as well.
Our fielding data is based on Brian’s own play-by-play system, and we’re also adding together all of this data to give you both actual (for past seasons) and projected wins above replacement (WAR), so that you can see just how valuable each player has been in the past, and how much value we expect him to have in the future. All of these are things you won’t find anywhere else.
Most importantly, these are not just some static forecasts that will lose all value once season the starts. Instead, we will be providing complete updates of our projections every week, taking into account every single thing that has happened on the field until that point. Wondering if a player will keep up his hot start or bust out of his slump? There’ll be no need to play guessing games anymore. We’re not implementing any shortcuts, either: Each update will use the same projection engine that we use in the off-season.
And to give our projections context, we’ve also partnered with the fabulous THT Fantasy staff to provide weekly updated depth chart projections. So you not only get a computer-generated projection for each player but also a forecast that incorporates the player’s actual expected playing time to tell you how much he’ll actually be contributing at the major league level. And again, the depth charts will be updated every week to make sure you have the freshest information possible.
We’re still not done, though. We’ve recruited 30 of the best bloggers (list below) from around the internet to provide us with commentary on over 1,300 players—and once the season starts, they’ll be providing fresh remarks so that you not only get our objective forecast, but the kind of subjective knowledge you can only get from people who watch these players day-in and day-out.
If you’re a baseball fan (and reading this site pretty much automatically means you are), you will love THT Forecasts, and with so much stuff (and weekly updates!), you will never get bored of it. And I say this without mentioning all the cool stuff we’ll be adding over the coming days, weeks, and months. All of this you can purchase right now for just $14.95.
We know that in a recession, money is tight, but that is why we’re offering the lowest price we possibly can. The price will go up next year (especially considering that we’ll be launching the 2011 version of THT Forecasts quite a bit earlier than this season), but if you purchase this year, we’ll keep the purchase price the same for you next year as well. That’s almost as good a value as Albert Pujols!
Just to recap, here is a partial list of what you will get with THT Forecasts:
{exp:list_maker}Oliver projections for the next six years for over 7,000 major and minor league players. These projections include hitting, pitching, and fielding statistics (the latter based on Brian Cartwright’s own play-by-play system), wins above replacement (WAR) projections, and coming soon, base running as well. You can read more about Oliver here.
Raw statistics for the past three years, including all the statistical categories listed above.
Major league equivalencies (MLEs) for the past three seasons, so you can see not just a player’s raw past statistics, but also how his numbers look adjusted for context.
Depth chart projections to tell you just how much of an impact a player will make at the major league level this season.
Over 1,300 player comments (and counting) from the best team bloggers on the internet, to give you a more subjective look at just about every player that matters.
And all of the above, updated each and every week, from now until October.
Plus, many more features, such as projected standings and personal player watch lists, with more to come very soon.
{/exp:list_maker}
All of this is available for just $14.95. So subscribe now! Or, subscribe to THT Forecasts and receive the Hardball Times Annual 2011 as soon as it is released this fall for just $34.95 ($39.95 for Canadian residents).
If you’re not convinced yet, we’ve made all the New York Yankee pages free for your perusal so that you can take a look at what we’re offering before you subscribe. It’s a sneak peak we know no baseball fan will be able to resist.
References & Resources
The player comment authors are,
Baltimore Orioles, Stacey Long, Camden Chat
Boston Red Sox, Evan Brunell, The Hardball Times
New York Yankees, SG, Replacement Level Yankees Weblog
Tampa Bay Devil Rays, Cork Gaines, Rays Index
Toronto Blue Jays, Jonathan Hale, The Hardball Times
Chicago White Sox, Mike Pindelski
Cleveland Indians, Ryan Richards, Let’s Go Tribe
Detroit Tigers, Brian Borawski, Tiger Blog
Kansas City Royals, Bradford Doolittle, Sports Radio KC
Minnesota Twins, Parker Hageman, Over the Baggy
Los Angeles Angels of Anaheim, Sean Smith, Anaheim Angels All the Way
Oakland Athletics, Sal Baxamusa, The Hardball Times
Seattle Mariners, Jeff Sullivan, Lookout Landing
Texas Rangers, Scott Lucas, The Ranger Rundown
Atlanta Braves, John Beamer, The Hardball Times
Florida Marlins, Alex Carver
New York Mets, Eric Simon, Amazin’ Avenue
Philadelphia Phillies, Corey Seidman, Phillies Nation
Washington Nationals, Chris Needham, Capitol Punishment
Chicago Cubs, Harry Pavlidis, Cubs f/x
Cincinnati Reds, Justin Inaz, Red Reporter
Houston Astros, Lisa Gray, The Astros Dugout
Milwaukee Brewers, Eric Johnson, Brew Crew Ball
Pittsburgh Pirates, Pat Lackey, Where Have You Gone, Andy Van Slyke
St. Louis Cardinals, Larry Borowsky, Viva El Birdos
Arizona Diamondbacks, Jim McLennan, Arizona Snake Pit
Colorado Rockies, Brandi Griffin, Purple Row
Los Angeles Dodgers, Eric Stephen, True Blue LA
San Diego Padres, Geoff Young, Ducksnorts
San Francisco Giants, Steve Treder, The Hardball Times
I’m definitely intrigued by the 6 year forecasts for minor leaguers. However, does Oliver have its own Superman Jr problem?
Here’s what Mike Fast wrote about PECOTA just the other day:
“However, as I was browsing through the Yankee hitter PECOTA cards that are freely available, one particular projection caught my eye. Baseball Prospectus projects Jesus Montero to hit .291/.334/.481 in the major leagues in 2010. Superman Jr., anyone?
The current leader among projection systems, Sean Smith’s CHONE, projects Montero to hit a much more subdued .255/.296/.425, and Dan Szymborski’s ZiPS projects a .273/.315/.416 batting line for Montero.
What could be causing PECOTA’s optimism about Montero? It looks to me like it’s the same thing that caused PECOTA’s interesting optimism about Wieters last year. According to Montero’s PECOTA card, his .317/.370/.539 batting line at Double-A Trenton was equivalent to a .312/.354/.529 batting line in the major leagues. That doesn’t seem right, does it? Aren’t the majors harder than Double-A? A lot harder?”
Oliver’s projection for Montero is 315/362/547!
Actually, hold on. Those are labeled Major League Equivalencies not projections. But Oliver keeps Montero’s stats going up and up. In what would have to be his MLB years his OPS is well over 1000. These are astounding projections.
Colin and Mike have pointed to the MLEs from the Eastern League as a potential source of the Weiters/Montero problems with PECOTA. Mike doesn’t buy PECOTA’s MLE for Montero of 312/354/529. And I agree, that doesn’t pass the smell test to me.
Oliver has Montero’s Trenton MLE at 287/356/543. It’s actually higher!
The Yankees don’t have any other big time hitting prospects to cross check, but Brian mentioned some very high projections for Heyward in his intro article.
Brian mentioned that he’ll be rolling out a lot of tests for prior year accuracy. Presumably those tests will be rolled out after many people spend their $15. Considering that I’m most interested in the minor league side of these forecasts, I’m not sure that I’m willing to shell out the money based on what I’ve seen of the Montero and Heyward projections.
I know Brian does very good work and is someone that imo does have the stature to say “trust me” and “this will be publicly justified with rigorous open testing”. But (you knew there’d be a but!) I have to say that the Montero projection shocked me and completely stopped me in my tracks.
Some things are very wrong here. Jeter’s OBPs for past seasons are slightly off from MLBs official stats. Teixera’s 2010 projections are different in the 2010 updating section and the Six year forecast section. The fielding forecasts stay the same for all six years, which seems unrealistic. Jeter’s fielding runs for 2009 are shown as 22.6 which seems way out of line with any other fielding metric.
As far as format goes, I find Fangraphs breakdown of individual WAR components to be much more helpful than what is shown here. With Fangraphs if you disagree with a single component of WAR you can substitute your own estimate, here you cannot. Didn’t you Beta test or have anybody vet this before you rolled it out as a for pay product? As it stands now it seems to be very buggy. To have confidence in projections one not only has to have confidence in the person doing the projections, which Brian’s track record with MLB player’s offensive projection inspires, but one also has to have confidence that the whole system has to be free of incidental mistakes, which doesn’t seem to be the case with this implementation. Plus Brian’s defensive numbers seem to be out of line with other defensive metrics and he needs to present his defensive metric for a full review for anyone to have confidence that his numbers have any validity where they differ from the others.
Hey Philly,
I agree that the Montero projection looks high, and this is definitely something Brian will be addressing. We plan to be 100% open about every possible issue the projections might have, and if we can’t find any problems, we’ll tell you what justifies a given projection. With that said, check out Brian’s article on his MLE system here:
http://www.baseballprospectus.com/article.php?articleid=8887
His way of doing the calculations appears to be the most accurate, by a good bit.
Hey Peter,
Texeira’s rate projections are the same in both sections—it’s only the counting numbers which are different because our depth chart editors predict he’ll have a little less playing time than Oliver says. There’s nothing wrong there. As for Jeter’s OBP, looks like the difference is that Brian is including SF in the denominator (i.e. he counts all PA), while MLB does not.
Brian will be writing up a description of his fielding system in the coming weeks, so you’ll get to see the inner workings for yourself. The defensive numbers should not be static from year-to-year, though—we will fix that ASAP.
By the way, I’m keeping a list of known issues and updates that we are working on right here, for maximum transparency:
http://www.hardballtimes.com/main/forecasts/todo/
We’ll add a link to the THT Forecasts page later today.
This is great, love the new format.
I’d also love to be able to download a csv with current playing time projections and rest of year forecasts on it. Just putting it out there..
-j
Hey Justin,
We’re working on getting that done as soon as possible.
Woot!
Your numbers do not put Montero in the company of Cabrera and Delgado. I’m looking at 351/400/676 at age 24, OPS over 1000 from age 22 to 25. That’s in the company of Pujols. Delgado had one season where he was as great as Montero’s projection, and only twice went over a 1000 OPS. As great as Miggy is, he has yet to do so (but close with .998 in 2006).
Delgado did not establish himself as a major league slugger until age 24. Cabrera did at 21, but his age 20 season was 268/325/468, far under Montero’s projected 2010.
My projection doesn’t look like much for 2010, but it is extremely impressive for a 20 year old. Applying the aging factors I get a peak at 312/365/560 at age 27. I too, am confident that he will rake. I just am not ready to accept as a 50/50 proposition that he rakes like Pujols and A-Rod.
Just an FYI, the projected standings have the average team winning about 79.5 games. Big fan of the to-do page.
Thanks, James. Not sure how that could happen. The logic in the code is set so that it all adds up. Hm…
Some comments re Wieters and Montero
Going into 2009, Witers had three years of college and one year of pro ball. I don’t know how much PECOTA uses college stats.
In 2009 Wieters had some very favorable park factors, Bowie 1.16L/1.07R for HR, Frederick 1.35L/1.57R, and he was already 23 coming into 2009. Most of a player’s growth is by age 21. I have similar curves to what Jeff Sackmann wrote here at THT last week using college data, but not high as his numbers.
My one year projections
BA OB SA wOBA WAR
2006 Fr 260/335/392 321 1.1
2007 So 261/345/407 333 1.7
2008 Jr 270/356/424 345 2.4
2009 298/387/485 380 5.0
2010 296/372/466 366 4.2</pre>
On the other hand, Montero is only 20 coming into 2010, so has more growth potential, and he’s always played in pitcher’s parks Tampa 0.81R for HR, Charleston 0.65R, Trenton 0.85R, and he’s moving into (eventually) New Yankess Stadium, which after one year is 1.34R for HR.
I am sure Montero will rake. My high numbers put him in company with Miguel Cabrera or Carlos Delgado in their early 20’s. He is on everyone’s top 5 list. As a Pirate fan, I am disappointed by the peak projections for guys like Alvarez or McCutchen, so not everyone projects high.
I will be looking into blending in a player’s actual track record of growth or decline the past few seasons with the standard aging curves to idetify those who aren’t growing or slowing down as expected.
There’s a computation problem with the cather’s fielding numbers. I work out of town (away from my home) and have a set of tables at home and at work. The home numbers the past weekend looked good, ranging from +7 to -15, with the guys I expected at the top and bottom. I sent in my last update from work, and I didn’t see that the steals vs catchers was messed up. I am rerunning them from scratch right now.
re Jeter – I have his range numbers -9, -12, -12 and -6 runs from 2005 to 2008, but in 2009 it goes to +18 (plus his usual +4 for hands and DPs).
For range, I first count how many balls a fielder gets to, then count how many ground balls went to the outfield for hits and proportion those between the infielders base on location. Once apportioned, those are added to the infield grounders to get total opportunities, into which is divided the apportioned ground ball hits. WOWY is used to compare each player with every other player in each park and by bathand.
Jeter’s expected GBH% was .247, .246, .242, .247 and .257 the past five years. His observed GBH% was .276, .291, .287, .274 and in 2009 .167.
It had been written that Jeter had a better year in 2009. My numbers have him below average in range for the previous four, but then quite good last year, and higher than other systems have said.
I will look at this more, but unless there’s a systematic problem with many players, it may be a one year fluke with Jeter. None the less, I am always thinking of ways to improve.
while im a huge fan of tht some of the yankee projections seem way off base which is why i wont be purchasing the forecaster.
for example, as has previously been pointed out, jesus montero will probably not hit 350 with 40 hrs.
also, why will joba chamberlain get worse and worse as he reaches his prime: never pitching more than 130 innings with eras balooning into the 5s by age 28. cc ssabathias eras also grow to an average level in three years.
????
Hey Steve,
I agree on Montero—we had a bug in the way we were calculating MLEs for minor leaguers and we’re working hard on correcting. Actually, we’ve more or less already figured out the fix and Montero’s projection comes down a lot. I’ll post something when we update the projections on the site in the coming days. As for Chamberlain and Sabathia, you should realize that pitchers tend to age very poorly. For example, take a look here at the aging curve that Phil Birnbaum found for pitchers:
http://4.bp.blogspot.com/_dUvDAtaIzO0/SJ_INTR5OvI/AAAAAAAAABE/dzjZ93MULMg/s320/graph.jpg
And here’s what Tom Tango found years ago:
http://www.tangotiger.net/adjacentPitching.html
The Chamberlain and Sabathia projections match sabermetric conventional wisdom, which should make you feel better about our projections, not worse. I’m glad you are a fan of THT, and I hope you give THT Forecasts another look.
Can’t wait for Baseball less than a month away. Early pick for world series: NYY vs STL
Does THT Forecasts come with a tool or software program for use in fantasy baseball drafts/auctions? I’m looking for something different than BP’s (mostly since 2 of the other 7 in my deep AL-only league will be using it this year), but I’m a fan of their application that gives you auction values based on customized league parameters.
Is anything like that available, or in the works for, THT Forecasts? thanks.
Hey Rob,
We do not have such a tool available yet, though it is on the extended to-do list. It won’t be available any time too soon, though. Still, I urge you to take a look at what THT Forecasts has to offer—weekly updated projections, depth charts, and much more. There’s no fantasy manager yet, but we’re also doing a lot of things that other sites don’t.
Rob, if you want a free auction value calculator, take a look at Mays Copeland’s site: http://www.lastplayerpicked.com/priceguide/
You can use the CHONE, ZiPS, etc., projections that he has loaded, or you can upload your own set of projections from a CSV file. It’s very easy to use.
Can we get sortable 6-year projection views on the leader boards and/or a downloadable spreadsheet that includes the 6-year projections? Something so you don’t need to click on each guy individually to compare 6-year projections?
Hey RFK,
We’ll try to add that feature as soon as possible.
As with early iterations of Pecota, the six-year projections almost always seem to decline from year 1-6…even with players currently in their early 20s. This does not resonate with me; is the aging curve projection truly right on this? Examples of guys where this seems strange are Jay Bruce, BJ Upton, Asdrubal Cabrera. Those are randomly picked to demonstrate the question, not outliers.
Justin Upton projects to be almost the same over that period, which also seems to suggest an unlikely aging curve assumption.
Note that this doesn’t seem to be the case with top prospects, but does with guys already in MLB. Hypotheses would be that there’s something hinky in the aging assumptions, but something that corrects for that in MLE adjustments.
Seems to me the value of six-year projections is pretty limited if effectively it’s an age-unadjusted phased decline…
Hey Bhs,
The age adjustments are fine—the thing is that Oliver identifies an earlier peak for most players than other studies have found. That’s why the guys you mentioned tend to decline between year 1 and 6—they have more decline years in remaining according to Oliver than years in which they’ll still be improving. If you look at younger players still in the major leagues (i.e. Justin Upton instead of BJ), you’ll see they still have plenty of improvement forecasted.
Dave,
I may be missing something but isnt there a flaw in tango’s analysis? the reason why strikeouts will peak at 20 is not because pitchers peak at 20. rather its due to the fact that any pitcher who pitches substantial innings at 20 will be super talented and thus strike out a lot of batters. at 27, there will be much more mediocre pitchers in the league who will dilute the strikeout numbers of the superstars. it seems to me that tango needs to follow a control group of pitchers rather than all the pitchers in the league.
thank you
Hey Steve,
That’s not how the aging curves work. The aging curves are built by comparing how a player did at one age to how he performed the next season. So it doesn’t matter how talented a pitcher is—if he strikes out 10 batters per game one year and 10.5 the next, the aging curve will be the same as if he strikes out 5 batters per game and 5.25 the next.
bhs,
The aging curves are based on millions of plate appearances from all levels of baseball. The reason you see most players show decline is that the peak age Brian Cartwright has found in his testing is 24—somewhat younger than what most people use (26-27). The difference is that Brian uses minor league data in addition to major league to construct the Oliver aging curves; as far as I know, he’s the first to ever do that. Using only major league data results in a bias as players that are not good enough to stick around in the majors drop out of the sample, making the peak seem higher than it actually is.
As for your last comment, what is relevant is our most likely expectation for each player. You’re not going to pay a player because you think he’s a better bet than someone else—you’re going to pay him for what you expect him to produce. I.e., of all the players currently in the MLB, maybe Jason Heyward is the most likely to be the best outfielder in 30 years. You’re still not going to sign him to a 30 year deal, since odds are a 50 year-old Jason Heyward will not contribute anything at all.
“If you look at younger players still in the major leagues (i.e. Justin Upton instead of BJ), you’ll see they still have plenty of improvement forecasted”
No, I cited several young players and looked at a large number of others and that actually doesn’t hold true generally. Even Justin Upton, the easiest case in MLB, barely shows any improvement after year one, in fact.
The issue is not that most players are worse in year 6 than in year 1; it’s that for a great majority of players they are worse every year after year 1 no matter their age is in year 1. Unless you’ve set the aging curve really low (and presumably erratically so) that wouldn’t be the outcome, I don’t believe.
When you say the aging curves are fine, what’s the basis for that, exactly? I’ve browsed a lot of guys and it looks like it is systematically wrong to me. You have the DB, if you have something showing otherwise would be curious to see it.
I can imagine a NPV approach where everyone is a poor bet to maintain value, but I’m not sure that has any value to anyone—-it’s relative value and risk that are relevant, isn’t it?
David,
Can we look at the aging curves you are using? Or are they a secret?
David, thanks for the response; I look forward to seeing what else THT has to offer.
And Mike, thanks for the link to the Last Player Picked tool. I imagine it should be fairly easy to load the THT projections into it.
RFK,
At some point, Brian will publish a whole series of articles explaining how Oliver works, and presumably he will cover the aging curves as part of that.
“As for your last comment, what is relevant is our most likely expectation for each player. You’re not going to pay a player because you think he’s a better bet than someone else”
I don’t think so, actually. I think the entire concept of player valuation is premised on the idea that player values are relative, not absolute. That’s why there is a baseline for both production and salary in every model, isn’t it? I think perhaps you’re confusing ‘value’ and ‘projected statistics’ and those are very different things.
On the last part of the above, if you’re saying that risk isn’t part of the value calculus, I believe that’s simply wrong. As would be the idea that risk is the same for every player.
So, I don’t see the relevance of a 30 year contract for Jason Heyward—-though I agree it would be a very bad idea.
On the peak question, I’m very curious to see the evidence on a 24 year old peak; that would indeed be very different.