It’s time for THT Forecasts

Today, we’re launching the 2011 version of THT Forecasts, the next best thing to a crystal ball for understanding what’s ahead in the 2011 baseball season. Forecasts gives you weekly-updated forecasts courtesy of Oliver throughout the season, as well as weekly-updated playing time projections provided by our depth chart editors.

In bullet-point form, here are some of the things you’ll find on Forecasts:

{exp:list_maker}Oliver projections for the next six years for more than 9,000 major and minor league players. These forecasts include hitting, pitching and fielding statistics (the latter based on Brian Cartwright’s own play-by-play system), as well as wins above replacement (WAR) projections. You can read more about Oliver here.
Raw statistics for the past four years, including all the statistical categories listed above.
Major league equivalencies (MLEs) for the past four 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 how much impact a player will make at the major league level this season.
More than 1,300 player comments from the best team bloggers on the internet, to give you a more subjective look at just about every player who matters.
A player watch list feature, so you can keep track of every player you care about on one page.
A fantasy value calculator that lets you set your league parameters, and tells you exactly how much each player is worth in your league.
Projected standings.
And all of the above, updated each and every week, from now until October.
{/exp:list_maker}
If you’re a fantasy player or just a baseball fan, there is no single better tool for understanding what’s to come than THT Forecasts. And best of all, it’s available for $14.95. If you’re ready to subscribe already, click this link. If not, feel free to browse around and see just how much Forecasts has to offer. We’ve made the player cards for the World Series champion San Francisco Giants free for your perusal, so you can take a look at what we have to offer before you subscribe.

It’s a sneak peek we know no baseball fan will be able to resist.

Using THT Forecasts

THT Forecasts offers a lot of features, so let’s take a tour around some of the most important ones.

Projections can be accessed in a lot of different ways on Forecasts. You can, for example, search for a player by typing in his last name in the search box on the left sidebar. If you type in “Greinke” and hit enter, you’ll be taken to a page that looks like this, where you can now click “View” to look at Zack Greinke’s player card.

At the top of the card, you’ll find our 2011 projection for Greinke, adjusted for his expected playing time. Since the season hasn’t started yet, you can see that his year-to-date numbers are all at zero, but overall we expect him to pitch 200 innings, going 13-8 with a 3.48 ERA*. Once the year starts, those numbers will start to change, and Forecasts will tell you both what we expect Greinke to do for the rest of the season, and what year-end line that should lead to.

*This article was written about a week ago, so some numbers referenced in it have since changed, and also of course Greinke’s projections will change when we update his team next week. We’ll have to update his player comment as well. Clearly, the Royals traded him at an inopportune time for THT Forecasts.

The next section on Greinke’s player card is his six-year Oliver forecast. This is a computer-generated forecast with no adjustments for major league playing time estimates. With Greinke, you can see that the computer is a little more optimistic about his playing time than our depth chart editors, penciling him in for 218 innings pitched in 2011.

The Oliver forecast is particularly good for two things: (1) You can see not just where a player is now, but how we expect him to develop over the next six years (with Greinke, for example, you can see that he’s projected to lose quite a bit of value between 2011 and 2016, going from 4.7 to 1.9 WAR), and (2) For players with no or minimal projected major league playing time, it gives you an idea of where they would be with roughly a full season’s worth of at-bats, rather than penciling them in for zeroes across the board.

After the Oliver forecast, we have a comment on Greinke, contributed by the fantastic Jeff Zimmerman, who warns fantasy owners to remember that with the Royals anemic offense, it doesn’t matter how good Greinke is—he’s still unlikely to contribute very many wins. (Note that player comments won’t be available for another week or two. We’re hard at work editing them right now.)

Following the player comment, we have three years worth of raw stats for Greinke. You can see how he progressed from merely very good to great before falling back last season. If you want to understand where his projection comes from, this is a good start. Better yet might be the Major League Equivalencies (MLEs) that follow. They’re not super useful for major league players, but for minor leaguers, they help put in context minor league statistics. Mike Moustakas put up some fantastic numbers in the minors last year, but as his MLEs show, that was equivalent to a .340 wOBA in the major leagues—slightly above average, but leaving much room for growth.

So that’s one way to use THT Forecasts. But there are many others as well. For example, I generally end up using the sortable leader boards instead. In the left column on the front page, you’ll find links to sortable batting and pitching statistics. Here’s what happens if you click on the “Sortable Batting” link.

The first page you are taken to is the “Rest of the Year” forecast leader board. This is exactly what it sounds like: These are our projections for the rest of the season, based on our playing time estimates as well as the Oliver projections. Right now, these will be the same as our “Full Year Forecast,” but once the season starts, this becomes an extremely valuable view for fantasy players looking to understand a player’s value for the rest of the year. Minor leaguers expected to get called up midseason will move up the rankings, while injured major leaguers will drop down.

This isn’t the only sortable view we offer, though. You can also choose to view “Year to Date” leader boards, which are pretty self-explanatory, a “Full Year Forecast,” which combines the “Rest of the Year” projections with “Year to Date” stats, and the “Oliver Forecast,” which is the computer forecast unadjusted for projected playing time.

What’s really cool about all these leader boards, however, is all the sorting and filtering options. Subscribers can filter the projections by position, league, organization and class. So let’s you’re a fantasy player in an NL-only league looking at catchers. You can choose to look at only National League catchers, and then, if you’re specifically looking for power, you can sort by home runs and see that we project Brian McCann to lead the NL with 22 homers. Or, if you’re a Red Sox fan, you can choose to select all Red Sox players, sort them by WAR, and find that Kevin Youkilis is Oliver’s favorite Red Sox, projected to accumulate 3.4 WAR in 2011.

A Hardball Times Update
Goodbye for now.

Better yet, if you’re one who likes to play around with numbers, you can download any leader board that you create into Excel by clicking the “Spreadsheet (CSV)” button. That means you can download all the hitter projections in one click, or just the projections for Orioles first basemen in Double-A. Feel free to play around with them to your heart’s content.

Like a good infomercial, however, we’re not done yet. There’s another option for viewing projections, and it’s one that I find myself using very often (available for no extra charge if you call in the next 15 minutes!). That’s the “My Forecasts” page, and it carries updated full year major league projections for only those players you care about. If you want to add a player to your watch list, all you have to do is click the “Save this player to your Player Watch list” button near the top of each player card. Your watch list is then accessible from any page on THT Forecasts—just hit the “My Forecasts” button, and you’ll have projections for all the players you care about on one page. I follow my favorite Red Sox players as well some top prospects, but for fantasy players, this is an extremely useful feature for keeping track of your fantasy team.

There’s one other feature I want to highlight in this tutorial, and that’s our fantasy value calculator. For fantasy players, there is no better tool on the web for valuing players year-round. Near the top of every Forecasts page, you’ll find a “Fantasy Price Guides” link. It takes you to a page titled “My Price Guides.” Click the “Create New Price Guide” button, and it will take you to a page that looks something like this. The page allows you to input your league-specific settings—number of teams, league, positions, positional eligibility requirements, statistical categories, and so forth. Then, all you have to do is hit “Save,” and the calculator will instantly give you dollar values for every player in baseball based on our projected rest of the season forecasts.

Not only is this tool going to be hugely helpful with your fantasy draft, but it will prove indispensable in-season as well. Since our forecasts and playing time estimates are updated weekly, you’ll be able to calculate rest-of-the-season fantasy values all year long, and that’ll give you a leg up when trading with your league mates and making waiver claims. No longer will you need to rely on outdated values, or try to guesstimate them based on updated projections. All you’ll have to do is press a button.

Moreover, you are not limited to creating just one price guide—you can create as many as you want and even edit them, and every price guide you create will be saved to your profile. If you play in five different fantasy leagues with five different sets of rules, you’ll be able to prepare for all five drafts with a minimum of hassle.

I hope that with this tutorial, you’ll be able to navigate THT Forecasts with a minimum of hassle. And as always, if something doesn’t make sense to you, feel free to e-mail our one-man customer service department (me) at
.

A final note

If you’ve read this far, I hope you’ve already subscribed to THT Forecasts after seeing all it has to offer. If not, though, I’ll make one more plea. The Hardball Times offers something on the order of five articles a day for free, but that doesn’t mean there isn’t a cost to running the site. Besides things like server costs and licensing fees, we like to pay the writers a little something as well. We don’t charge for articles, and we believe in getting paid only for value-added content. If you’re on the fence about THT Forecasts, please consider that by subscribing you’re helping to support The Hardball Times, ensuring that we’ll be around for awhile to provide you with great baseball content five days a week, 52 weeks a year.


25 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Michael Bush
13 years ago

I am rather disappointed with the Oliver forecasts.

I have looked at the Reds players who I am pretty familiar with.  The young players (21 -24) don’t show any growth as they grow into their next 6 years.  Certainly this goes against all player development research.  Some won’t develop but most will show improvement.

At this point, I like the card but the Oliver forecasts as they currently exist are worthless and I feel like my money was taken.

robert.twining
13 years ago

Good product and interesting layout.  I am missing the Washington Nationals.

Peter Jensen
13 years ago

A big difference between Brian’s research and others is that he has researched players at many levels, not just those who stuck in the majors.

Why would anyone study the aging of non major league players to predict the aging curves of players who reach the major leagues?  That is a bigger,but similar, mistake to Bradbury basing aging curves on those major leaguers who were still playing in their late 30s.  The main reason minor leaguers never achieve a major league career is that they DON’T keep improving after their mid twenties and those that get promoted to the major leagues and stay there DO.  My best year as a baseball player was when I was 12.  Beware when Brian starts including MLEs for little leaguers.

Dave Studeman
13 years ago

Robert, looks like there’s a bug for the Nationals—very embarrassing.  We’ll get right on that.

Peter, I’ll ask Brian to drop by to discuss his aging methods.

Lee Panas
13 years ago

I signed up and received an e-mail confirming access.  However, every time I try to view something, it says I don’t have access to that page.  Is there something wrong with my account or does it take a while for the registration to kick in?

Lee

Dave Studeman
13 years ago

Sorry you feel that way, Michael.  Brian’s research has found that batters peak at age 25, not later, and that they tend to stay flat between 23 and 27. A big difference between Brian’s research and others is that he has researched players at many levels, not just those who stuck in the majors.

We don’t want anyone to feel that we took their money.  Please email me (
) and we’ll refund your subscription.

David Gassko
13 years ago

Robert,

The Nationals things is fixed.

Peter,

There’s no right answer here. Using only major league statistics biases aging curves in the same way that Bradbury’s study does, as it selects on players who were performed well enough to stay in the major leagues (and therefore were to some degree lucky). Including minor league statistics helps remove that bias, though it does introduce a potential problem if minor league players have different aging curves than major leaguers. We don’t know that they do, though.

Lee,

Could you please send me an e-mail? It looks like your payment may not have gone through. Include your confirmation e-mail please.

Toffer
13 years ago

Did THT ever get around to writing the articles that were promised to paying subscribers back in early March? Although I found that THT’s forecasts were pretty decent I would find it very difficult to subscribe again to a service that makes promises but then refuses to fulfil them (“Over the next few weeks…I will be writing articles which will examine some of the inner working of Oliver, measure its accuracy, and show leaders…I’ll do is test the accuracy of my projections for college players; that will be followed by a discussion of aging curves.” March 04, 2010 – http://www.hardballtimes.com/main/article/introducing-oliver/); and also blows off customers when they ask for those promised articles 3 months later and again are given a promise, “These are still planned, and I am scheduled for an article THIS Friday” (June 07, http://www.hardballtimes.com/main/fantasy/article/twisting-oliver-one-category-pitching-dynamos/) but again don’t follow through with the promised articles.

If THT makes promises about their product and fails to deliver is there any reason we should trust the projections themselves?

David Gassko
13 years ago

Toffer,

I’m sorry that you feel that way. Brian puts a lot of work into putting together the Oliver projections, and since he also has a real-world job, that doesn’t leave him much time to write articles. Hopefully, he’ll find the time to give readers as much detail on Oliver as possible, but I can’t make any guarantees for another person. The only guarantees I am in a position to make are for THT Forecasts itself (since I run the thing), and those are embedded within this article.

As for the Oliver projections, there have been a number of articles comparing projection accuracy that have found Oliver to be a very good system. You can find those with a bit of Googl-ing.

Mike R
13 years ago

I also am denied access to player pages.

David Gassko
13 years ago

Mike,

Please try clearing your cookies and closing your browser. If you still don’t have access, please send me an e-mail (
) and we’ll get it figured out right away.

Ducat2,

Thank you.

Peter Jensen
13 years ago

There’s no right answer here. Using only major league statistics biases aging curves in the same way that Bradbury’s study does, as it selects on players who were performed well enough to stay in the major leagues (and therefore were to some degree lucky). Including minor league statistics helps remove that bias, though it does introduce a potential problem if minor league players have different aging curves than major leaguers.

David – If you mean that aging curves are difficult because players age differently you are correct.  But if you mean that the methodology behind Brian’s aging curves has just as much chance of being correct as aging curves limited to major league players, you are wrong.  Bradbury studied a subset of all major league players and Brian is using a superset that goes beyond major leaguers.  Both provide an answer to SOME question, just not the question of how major league players performance changes with time.

But lets go beyond the aging question and look at other oddities of Brian’s projections.  I looked at Andrew Torres as a quick example. 

1)There is overwriting in the HTML code.  Why would you not proof this before release.

2)The team expert has Torres projected with 85% playing time, Brian at less playing time than last year.  That’s OK, there certainly is room for legitimate differences of opinion, but Brian’s multiyear projects has Torres playing time increasing every year through his age 38 year.  That must be a code mistake.  Code mistakes happen, but aren’t you supposed to be adding some editorial oversight.

3) Brian has Torres’s 2011 fielding runs at 19.6 at 467 PAs.  But he says that Torres’s fielding runs in 2010 were only 15.1 in 568 PAs.  According to Brian, Torres has NEVER fielded at a 21run/500PA rate in the last three years.  Why is he projecting him to improve his fielding at age 33 AND stay at that level every year after that through age 38?

4) How come there is no way to pull up fielding stats for all the players as there is for pitching and hitting stats?

5) Like Toffer I had asked some specific questions about Brian’s 2010 projections and was promised future explanations which never came.  You rushed those projections to make them available for fantasy players in time for their drafts.  So initial errors were somewhat excusable.  But now its a year later.  You say above that you “run the thing”.  I would think that would mean providing assistance to Brian in correcting mistakes and inconsistancies, and providing explanations and descriptions of the methodology where it is unclear and questionable.

Dave Studeman
13 years ago

Peter,

I’ll let David and Brian answer your questions, but I have a question and a comment for you:

– What HTML “overwriting” are you referring to?  I don’t know what that means, and I don’t see any problems on the page.

– You are in no position to criticize our process. You did not experience our process.  You can see the output, and you have legitimate questions about it. But when you ask “aren’t you supposed to be adding some editorial oversight?”, you’re just being mean and not helpful.

We want to have useful dialogue, provide better explanations and improve our process, but please refrain from trolling.

David Gassko
13 years ago

Peter,

Once more, I disagree on the aging curves. You seem to be conflating your opinions with fact—there is simply no one in the world who knows for sure what the correct aging curve for a major league player will look like. You say that Brian’s work on aging makes the same mistake as JC, but it’s actually you who is making that mistake. Looking only at major league statistics in creating an aging curve biases your results in much the same way as looking only at major league players with x amount of playing time, the only difference being that here x is 0 rather than 300 (or whatever JC does). Brian’s method more or less solves that problem, though it introduces the potential bias that minor league players might not have the same aging curves as guys who make it to the majors. Note the “might” in there—we have no way of knowing whether or not that’s true. Personally, I prefer the potential bias to the certain one, but certainly others can feel differently. It’s a question of opinion, though, not fact.

As for the rest of your points:

(1) I have no idea what this means, but the player card looks fine to me.

(2) You insist there’s a coding mistake, but in fact as far as I can tell, there isn’t. Instead, what Brian’s playing time projections do is use a player’s performance (in the previous season, I believe) as one of their inputs. So players who perform at a high level will tend to have more PA, all else being equal, and vice-versa. Oliver loves Torres’ defense, and so it believes that he deserves a lot of playing time. Since he’s starting at a low base, it assumes that teams will recognize this and his playing time will inch up.

(3) Again, you’re simply wrong. Torres’s MLE has him at +16.1 runs in just 205 AB in 2009. That would be a 32.3 run pace given his Oliver projected playing time for 2010. That information is right on the player card.

(4) The fielding stats are available for all players in the batting leader boards. It’s the column right before WAR. Since, as far as I can tell, you don’t subscribe, you would of course have no way of knowing this.

(5) I’m not sure what “mistakes and inconsistencies” it is that you are referring to—the mistakes in your comment all seem to be your own. I do agree that it would be very helpful to have an article from Brian giving more detail as to how Oliver works, and I have encouraged him to write one, but ultimately that depends on Brian’s schedule. Oliver is not my system, and so I am in no position to write anything about its details myself.

And finally, let me second Dave’s comment. I am more than happy to answer any questions that I can and look into any problems that might arise. However, there is a very bright line between that and trolling, and I have no interest in responding to those who wish to cross it.

Peter Jensen
13 years ago

Both Davids – I apologize if you thought I was trolling or offering mean criticism that was not meant to be helpful.  What Brian is trying to do with his projections is truely a massive job and he deserves credit for even attempting it.  When David G. said that he “ran the thing” in response to a previous post, I assumed that the Forecasts were a collaborative effort, and that he and perhaps others were providing critical feedback to Brian to make the Forecasts the best possible product for THT to sell to its readers. 

On my browser when I pull up individual player cards including Torres, the “projected percent of teams playing time at:” overwrites “Save this player to your Player Watch list” and the player’s birth date.  Perhaps this is just my system, but I can think of no reason why that would be so, and I certainly couldn’t have presumed that it was limited to me when I made my comment.

As far as aging curves goes, I feel David G. is very wrong in his explanation and that Brian made the incorrect decision to include minor league players, but we are not going to resolve our differences of opinion in this forum. 

Brian does have Torres MLE as 16.1 runs in 2009.  But if you look at the section above that he has Torres as having an actual fielding runs as 3.3 runs in 170 PAs. How that and his 2.6 runs in 65 minor league PAs equates to 31.3runs/500PA rate is not explained and seems unfathomable.

In the menu area you have sortable batting and sortable pitching but not sortable fielding.  My question refered to that omission.  If I were to become a customer I would like to be able to call up the fielding projections of all the team’s players on a single page.  Perhaps subscribers can do that in another way that I am not aware of.

I find it hard to believe that you are defending Brian’s static fielding projections for Torres from ages 33 to 38 no matter how high Oliver is on Torres true fielding talent.  Nor do I find your explanation of Brian’s playing time projections understandable.  You state “So players that perform at a high level will have more PA, all else being equal, and vice versa.”  But Brian is projecting Torres to lose about 40% of his war from 33 to 38.  How does that substantial decrease in performance equate to MORE playing time?  And its not just Torres.  Brian has Cody Ross projected with 550 PAs at age 35 after he has declined to minus WAR performance.  The number of players that follow this pattern is what caused me to conclude that it was due to a coding mistake.

Dave Studeman
13 years ago

Thanks for the feedback on the page, Peter.  We have tested the pages many ways and haven’t seen that.  What type browser, operating system do you use, and what is the size of your screen?

Does anyone else have “bunched up” code on the player pages?

Peter Jensen
13 years ago

Dave – Explorer, Vista, 21”.

David Gassko
13 years ago

Peter,

Rest assured that Brian is writing up some explanatory articles and hopefully those will be up soon. Obviously, I can’t make any promises, but he does want to give readers a better understanding of what he’s doing with Oliver and I’ve been pushing him as well.

As for the playing time projection, from what I understand (and I could be wrong), what Brian basically does is use some combination of a player’s weighted average playing time, his age, and his previous WAR. So the equation might look something like:

x*Weight_Avg + y*WAR + Age Adjustment

In that case, as the weighted average of PA for Torres goes up in the Oliver projections, that might counteract the age adjustment. Brian could definitely give you a better answer though.

The static fielding projections exist because I don’t believe Brian has been able to come up with a satisfactory aging curve for fielding. From personal experience, I can tell you that when I’ve tried to do the same, I’ve come up with a very flat curve. Seems wrong, but I don’t know what I could have done differently. Again, this is something that Brian could answer better.

Fielding projections are available in the sortable batting screen, so yes, subscribers can do exactly what you are describing.

Peter Jensen
13 years ago

Not to pick on you guys but here’s another question that has me perplexed.  Pablo Sandoval played the entire 2009 season for SFN. He hit 25 HRs in 572 ABs.  How come his MLE has him hitting 28 HRs in 571 ABs?

David Gassko
13 years ago

The AB thing you’d have to ask Brian about. As for HR, it’s likely because San Francisco is not a great HR park.

ducat2
13 years ago

The layout looks fantastic.  The effort the writers put into THT’s quality articles and features, all without compensation, is to be commended, not criticized.  I predict the $14.95 spent on the THT Forecasts will be well worthwhile.

Tom T
13 years ago

Really excited to view the projections, but I am also denied access to the page.

Just checked paypal and it indicates that payment was received. I’m sure it’s easily fixable, just really anxious to take a look at your projections! Thank you.

Toffer
13 years ago

David Gassko – I appreciate that the author has a full-time job but I just wish that THT would under-promise and over-deliver to paying customers rather than vice versa. Promising to publish an article in less than a week and then never doing so at all is slightly disrespectful, particularly since I had a fairly basic question that could have been answered easily.

Could you link to some of the comparisons? The one I followed (Tangotiger’s Forecaster’s Challenge) THT/Oliver was not submitted.

Either way, I did enjoy THT’s Forecasts last year and the additions look great but I really hope that we can get some more info about THT’s forecasts and its underlying assumptions, strengths, weaknesses, etc.

David Gassko
13 years ago

I think that’s fair, Toffer, and I personally try very hard to make sure that any commitments I do make I know will be fulfilled. I do hope that Brian will have something up soon that gives some more color as to how Oliver works.

Here are some links pertaining to Oliver’s accuracy:

http://www.insidethebook.com/ee/index.php/site/comments/evaluating_the_2009_forecasts_chone_zips_fantastics_win/#12

http://www.fangraphs.com/blogs/index.php/projection-vs-projection/

Brian Cartwright
13 years ago

I am concerned when people can find issues with our published results, and I am looking into these. I do have further work to do on play time estimates (although we have depth chart estimates for current season MLB players), and I need to check the three year weighting for defensive projections.

We’ve been having a discussion of the aging at Tango’s blog, and yesterday, in response to Mike Fast, I ran some numbers and posted a graph of the results. Further research today led to my finding a small adjustment which gets a resulting aging curve that I think most will be more comfortable with.
http://www.insidethebook.com/ee/index.php/site/comments/tht_forecasts/

Here’s what I just posted there

“Some people have questioned using minor league stats to profile major leaguers. In the early ages, almost all the players are in the minors and it’s the only sample available for those players 17-23, when the most rapid changes are occurring. Likewise, I can say that after 25, most of the players input to the age study are in the majors, and I didn’t think that mixing the minor and major data was going to make a substantial difference in the results, as I was comparing players in the same league in consecutive years.

Looking at the ‘Lateral’ graph, players who changed leagues (and thus were not included) but those leagues were at the same level, had a pre-25 growth very similar very similar to the league repeaters, but it was flat into the 30’s. Could I also use this data?

Where the two curves differed was in the post-peak years, when a large majority of the players were in MLB. There are differences in talent level between the NL and AL, but in many ways they are like one large league, and I was not including in the age study players who made lateral movements between the two major leagues.

When I adjusted my code to include these players, HRs peaked at 29, extremely flat 25-30. All categories carried skills better into the late 20’s. I have updated the spreadsheet linked in #34.

I will generate a composite line to see what a typical player’s stats would be at each age, calculating a wOBA which I now believe will peak closer to 27.

I actually feel very good to be wrong – I still believe my basic model was correct, and this adjustment, which does not violate any of my previous criteria, comes up with a result that is more intuitive and acceptable.”

https://spreadsheets.google.com/ccc?key=0Akieb136KCz2dEJJMDl2WVYxQV9JSEtWZ2t3MVc3MWc&hl=en&pli=1#gid=0