Oliver Projections
The Oliver projection system for batters is now available in the player pages, projections pages, and you can get a customized player list using the my projections feature.
Huge thanks to Brian Cartwright for allowing us to use his projections and you can read all about them at Statistically Speaking.
David Appelman is the creator of FanGraphs.
Yikes, .370 wOBA for Hanley? That seems extremely low…. When did he turn into Brian Roberts?
Yeah, they have pretty much every aspect of his game regressing mightily. I certainly hope this is not the case, I would be very sad.
His career OBP is .379, I haven’t known many players as toolsy as Hanley, who has shown improvement in each facet of his game peak at 24 and significantly regress at 25! Well, I am sure there are a few, but not with Hanley’s pedigree and just recent MLB success….
With a career .379 OBP, that .370 wOBA ‘projection’ just seems like a waste of time, sorry to say…
HanRam did not produce in the minors, that’s why its lower than you’d think.
No, but he was the Red Sox top prospect according to BA for like 4 years running before he was dealt… Every scout was in awe of him… But you’re right, outside of rookie ball (51 games) he was average in the minors…
Hey, Brian Roberts is 345
The past years weighting is causing the projections to lag a little behind someone like Ramirez who broke out to a very high level, but 370 still has him as the best offensive ss in the majors.
MLE PROJ
2002 18 BOS A- 419 367
2003 19 BOS A 311 331
2004 20 BOS AA 329 327
2005 21 BOS ML 295 314
2006 22 FLO ML 360 333
2007 23 FLO ML 404 357
2008 24 FLO ML 401 370
Yeah, I was referring to Hanley’s projection (.370) vs Brian Roberts 08 wOBA (.369)…
Awesome, I’ve been a fan of the Olivers this offseason.
I’m sure Brian will drop in at some point and answer some questions about certain players, but in general I believe his projections are Marcel + more stuff.
All I can really say is that this is the reason why we display multiple projection systems.
Definitely, its good to have more info! Thanks guys… I am just a big Hanley fan, seems a bit low balled IMO..
My projections are modeled on Marcel, weighting past seasons a decreasing amount. However, Marcel does only the last 3 yars, I do an unlimited. Marcel does not use minor league data, I do, calculating my own MLE factors.
Only 2 guys cracking a .400 OBP? Is it going out of style like AVG?
I don’t know about you all but I’m going after the second best HR-hitting catcher this year, Mike Wilson. Who knew that this 113-year-old still had something in the tank?
Whoops, that is the wrong Mike Wilson, should be fixed now along with one other mismatched player.
Is there any way to get the output of all three systems in one Excel file without having to go thru ALL the pages, cut/paste/forward/repeat?
Unfortunately there’s not, and it’s probably not going to happen. You can always download the spreadsheets from the original sites and do vlookups in excel. That’s probably the easiest way I can think of doing it.
Will pitchers be available as well? And Are these projections park-neutral or specific to the park you project them to play at next year?
Um, only 11 players out of 1453 are going to exceed .300 BA? That would be down from 33 in 2008, so that’s quite a slide in one year. Last year 44 players had SLG% over .500 but Oliver predicts only 28? Is there a defined statistical basis for the predicted lack of output?
Is this a computer glitch caused by the regression of offensive production we have seen since the elimination of PEDs? Is there anyway to explain to the computer that just because offensive output fell sharply it doesn’t mean the slide is progressing/progressive?
Ed, that’s the same with all projection systems. My numbers showed that batting average had a year-to-year correlation of only 0.422 since 1970, which means projections of hits are only based on player’s performance by 0.422, and its regression to the mean component is weighted at 0.578 (1- .422).
OK that I can understand, but where many have had some trouble is in trying to establish how we are supposed to use these numbers to handle our fantasy drafts. I take these projections as simply a way of ranking one player ahead of another, and not as a way of predicting a real world result (as it seems very very unlikely that offensive output at the top is likely to regress this much in one year). Would you agree with that?
Also I would ask how the changing of the game, and it essential components (the size, strength, of the players), change the year to year correlation? Is there something about the game of baseball that has changed so drastically since 1970 that a different starting point for calculating the correlation would be beneficial? Did the increased offensive outputs of the 1990s-early 2000s skewer some predictive systems and how are statisticians taking that into account (if at all)?