SubscribeAnother advocate of drafting pitchers in the first round: Baseball Prospectus’s Player Forecast Manager. Here’s an excerpt from an article I came across:
Q: It looks like the overvaluation of pitchers is back again. After the first day, the valuations were normal, but the PFM is currently showing pitchers as the second, third, fourth, and 6th-8th most valuable players in the game for fantasy purposes.
BP’s Ben Murphy explains that, although it might look like a bug, the high pitcher values are correct based on their methodology.
So why does everyone who runs the numbers come to the conclusion that the top-tier starters are worth taking in the first round? Is everyone overlooking the unpredictability factor?
Related posts:
Interesting. With standard league settings, and Pos Adj #1, you get 4 pitchers in the first round, with Rivera and Papelbon at the beginning of the 2nd. However, with Pos Adj 0 only Santana is ranked among the first 12.
I’m not 100% sure how you and others calculate fantasy dollars, but I’d bet that the reason that good pitchers are “overvalued” is because fantasy dollars are based on z-scores rather than “fantasy points earned.”
What you see is a lot of pitchers forecast to be about the same, with a couple aces above the pack and then a bunch of guys near replacement level.
Wins, ERA, and to some extent WHIP are fairly hard to accurately predict because there’s a lot of luck involved with earning those stats.
Add to that the fact that a top pitcher is only a 4-cat guy, compared to a han-ram or Sizemore who’s a 5-cat guy.
@Molson: I’m not sure I understand the distinction. What would be an example of basing dollars on “fantasy points earned?”
Hanley Ramirez (67 RBI last year) and Grady Sizemore (.268 AVG) are both four category players in my book. There doesn’t seem like there’s really been a consistent five category hitter since A-Rod’s run in the late 90’s. (Beltran was pretty good but didn’t quite have the AVG you would like…)
Hmmm, I would simply question the valuation system then. I use Todd Zola’s Replacement Level system from Mastersball.com and my pitchers are ranked fairly in line with where they get drafted. Johan mid 2nd round, along with Sabathia, then some in the 3rd, etc.
I think on the whole, pitchers ARE slightly undervalued when compared to ADP because so many owners try to wait on pitching regardless of what the dollar values say. I can’t imagine the BPro values being mathematically correct though. Maybe they aren’t using a typical 67/33 type split?
@Mike: I don’t subscribe to BP, but I wouldn’t be surprised if their split ends up close to two-thirds/one-thirds: When the Price Guide generates values for a standard league, it splits the money 66/34, and it also puts four pitchers in the first round. :-)
I’m at a real loss here. With CHONE projections, 3 pitchers project to earn over $40!! I have NEVER projected a pitcher to be worth $40 I don’t think. Then again, maybe Pedro in his heyday was projected that high. For me, Johan has typicall been projected in the mid-$30’s as the top pitcher. However this year, I have no pitcher projected for even $30! So something is seriously wrong here and I’m not sure what it is considering we both have similar hitting/pitching splits.
The standard deviation for CHONE’s xERA is really, really small. It comes out as 3.97 for a standard roto league. This number should be around 10. It’s equally bad for xWHIP: 7.4 vs. expected around 15.
So translating Santana’s value into an offensive player, the valuation system sees him as if he were a 110 R, 31 HR, 125 RBI, 10 SB, with a .314 average in 550 AB.
Looking at him like that, what would you pay for him? That line is pretty much what’s projected for Pujols or A-Rod, so I’d say the answer is “whatever you’d pay for them.” Worth it, probably not, but that’s the reason the price guide puts out those numbers.
I don’t think the SDs from the projections are “bad,” they are just referencing a different scale. The projections are more conservative than what will actually happen (and rightly so), but it makes sense that the SDs will have to be lower.
It just so happens that I’ve got a post going up on this tomorrow, so you are welcome to comment there as well. :-)
Well, in a way, the SDs from the projections ARE bad in that they don’t reflect reality. If a system projects every pitcher to have 10 wins except for Santana who it projects to have 20, is Santana’s win total really worth twice as much as all other pitchers?
Of course not. In those guys they predict for 10 wins, there will be folks who get 18 and folks who get 2. The variance will be large (hence a large actual SD), but you just can’t project who will get 18 and who will get 2 as it’s essentially random.
If the projections are extremely unreliable, you want to use a higher SD, not a lower one.
Take an extreme case as an example. Say we’ve got 21 players and give each a number from 0-20. Here, the mean is 10 and the standard deviation is 6.2.
If the numbers are assigned randomly, our best attempt at a prediction is to project 10 for each player. But the SD of that prediction would be zero. This means that all those players get infinite value from that stat! You want it such that all the players would get zero value from that stat.
Now, what if you knew who got 10 and could tell who was above or below average and that’s it. Then your best bet is to project the top half for 15 and the bottom half for 5. Here the SD is 5 and the top half would add 1 unit and the bottom half would subtract one unit. But it’s really worth less than that to get a random top half guy. But a random player in the top half isn’t 1 standard deviation above average. A random player in the top half is .81 standard deviations above average. We’re interested in how our team of players actually compares to the average once the season is over, and in this imaginary category we can expect to be .81 standard deviations above the average.
I subscribe to BP and the PFM allows you to customize your allocation of $ to hitters vs. pitchers.