If you had the first pick of your draft this year and you took Jose Reyes over Hanley Ramirez, you probably (definitely) aren’t very happy right now. But that doesn’t mean it wasn’t the right choice at the time. You made a choice to take a whole lot of speed with a nice splash of power thrown in over the more balanced power/speed combo. And even those descriptions, as difficult as they are to differentiate in terms of fantasy value, don’t tell the whole story. We have to worry about runs, RBIs, and batting average (and those are just the typical roto stats-if your league uses sac flies as a category, you’re really in trouble). While it is easy to look at two players’ homerun totals and be able to tell who will provide you with more power, it is considerably more difficult to weigh five stats and know which player has more value.
Plenty of fantasy analysts have tried to solve the five stat problem by coming up with what we call a valuation-a single number that tells us how much value a player has. This is very similar to VORP, value over replacement player, except it is designed for the fantasy player. If you know that Hanley is worth $38 and Jose is worth $34 then you probably want Hanley because he has more value.
There are some problems with valuations, however. One is that as you can see from the Jose/Hanley example, values are often given in monetary amounts. This is because so many people play in auction leagues… just no one you know. I’m half-kidding. If you are deep into fantasy, you may be more familiar with auctions, but the vast majority of players are more invested in snake drafts than in auctions. Unfortunately, if you’re not in an auction, these values look completely arbitrary. They tell you who is worth more but the actual number does not have any relevance to your fantasy league, the way VORP does to MLB. Ramirez isn’t worth 38 extra points in your league or make you 38 times more likely to win. He’s merely worth an arbitrary 38 dollars.
The other problem with valuations is that the people who create them don’t necessarily know what a player’s valuation should be any better than you do. I asked one person in the industry how his group determines the amounts they publish and he said, “We just choose what we think he is worth based on our experience.” Some of the better websites have a system for creating valuations but their information is proprietary and you have to rely on your own experience with their website to determine whether you trust their numbers.
Fed up with trying to determine whether a steal or a homerun is more valuable a few years ago (yes, a dinger is worth much more in real baseball, but in fantasy, maybe not so much), I set out to create a better system for evaluating fantasy players, which I like to call VOFP or value over fantasy player.
One problem with finding fantasy value is a lack of data (couldn’t help the pun). Some of you out there might be in 40 leagues but my four aren’t cutting it in terms of sample size. Fortunately, though, I play in the National Fantasy Baseball Championship (NFBC), which currently has 390 owners in its “main event”. Using the NFBC data, it is much easier to see how much specific numbers help you and therefore what players are worth.
VOFP can be used in the NFBC or in a regular Yahoo league, but to understand how VOFP is derived, you must first understand how the NFBC works. It is split up into 15 team leagues that are organized just like your regular roto league. The leagues have 14 hitters and nine pitchers with the usual ten stat categories. The difference is that along with your league competition you are also in an overall competition with all 390 teams that works the same way. If you have the most homeruns, you get 390 points. If you have the least, you get one. My goal every year is to finish in first place and win the $100,000. In order to do that I need to average around 40th place in each category or score 350 of the 390 possible points (yeah, I know, but it’s harder than it sounds).
The idea is that you need a certain amount in each statistical category in order to reach 40th place (or 350 points). Let’s take homeruns, for example. You’d need about 294 homers to land 40th place most years. That means you typically need to get 21 jacks from each of your position players. Some might give you 31 and some might give you 11, but the players need to average 21.
With this in mind, you can assign a point value to the number of long balls you get from each guy on your team. While there is a larger discrepancy at the extremes of the standings (the #1 guy had 19 more homeruns than the #2 guy last year), once you get to 40th place, point values stay more consistent. Testing this, I found that each homer usually moves you up and down the rankings by about four places.
Let’s take a look at how many points each stat is worth. Ideal is how many each player in your lineup has to average to ensure you of a highly successful season:
Category Ideal Points
Run 84 1.69
Homerun 21 4.08
RBI 81 1.66
Average .286 1.3
Steal 13 4.13
If you had 14 players who put up the “ideal” line at the end of the season in the NFBC (and you had pitchers who were equally good), you’d not only have a great chance of winning your league, you’d have a great chance of winning the $100,000. For smaller leagues, you’d have to get more out of each player, but you’d have to get proportionally more in each category. In other words, these values are applicable to most leagues (though not Hacking Mass!). For the purpose of finding values of players we’ll consider this our baseline. Players with this exact line have a VOFP of 0. So, really, rather than replacement players, we’re talking about ideal players because that is what you want your fantasy team to have. The metric would more appropriately be named VOIFP but that is more than a mouthful.
The points are how much a player’s VOFP moves up and down depending on his stats. For example, if you have a player with the ideal line exactly except that he had 22 homers, his VOFP would be 4.08 and he’d be worth four points more in the competition than an ideal player. On the other hand, if your guy only had 20 dingers, his VOFP would be -4.08.
Let’s use VOFP to explore this year’s PECOTA projections. We’ll look at players from a typical 2009 first round in the NFBC:
Pick Player PECOTA VOFP To Date # VOFP (prorated) 1 Hanley Ramirez 273 156 2 Jose Reyes 309 -30 3 Albert Pujols 250 382 4 David Wright 205 248 5 Miguel Cabrera 83 264 6 Grady Sizemore 146 21 7 Ryan Braun 170 149 8 Jimmy Rollins 100 -91 9 Ian Kinsler 74 305 10 Ryan Howard 75 146 11 Josh Hamilton 28 -80 12 Chase Utley 86 191 13 Mark Teixeira 25 173 14 Carlos Beltran 121 273 15 BJ Upton 35 -109
The ‘VOFP to date’ is for entertainment purposes only. The sample is too small for us to draw any firm conclusions, except to give a big woot, woot to Pujols for being so consistently great. Looking at the PECOTA values, however, can be very informative. For one, we have an answer to our initial question: it seems like if we put our faith in PECOTA, Reyes was actually more valuable than Ramirez. Both were good bets, but Hanley seemed like the bigger health risk at the time, so the line of thought that said Hanley was too valuable to pass up for Reyes was probably borne of the theory that homeruns are far more valuable than stolen bases, which we now know is incorrect, based on the first table.
We also see that Cabrera, Hamilton, and Teixeira were likely overvalued, while you’d be better off with Beltran-or Soriano, who didn’t even make the list but had a VOFP of 127. Trading for these guys over someone higher on the list right now-like Cabrera who won’t hit .377 all year-is probably a good idea.
While we are on the topic of Cabrera, let’s see why he had such a low VOFP. PECOTA had him with 94 runs, 32 homeruns, and 111 RBIs, good for 112 VOFP just from those three stats. A .294 batting average, weighted for the number of at bats, gains him another 12 points. So how does a guy with 124 VOFP end up with 83? Simple. He doesn’t steal bases. With only two projected bags, he loses 41 points. It is easy to think we can draft speed elsewhere, but why avoid it now when you can get similar numbers from Beltran along with the steals at a later draft position? It helps your team a lot more in the coming rounds if those swipes are already in the bank.
There is an almost unlimited amount of information we can look at with VOFP. In the future, I’d like to discuss VOFP for pitchers, VOFP and injuries, VOFP alterations for different size leagues, VOFP and positional scarcity, VOFP vs. Godzilla. For now, it will have to be enough to learn that VOFP tells us steals are worth a whole lot in fantasy, Ramirez and Reyes aren’t much different, first round draft picks don’t always progressively lose value, and Pujols is the only player who is consistently a god.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now
Zero is just the floor for an ideal player. Guys who get 21 homers score zero, but guys who get bigger numbers score well over that. Again, using Cabrera, he was projected at 32 homeruns, so those 11 extra dingers translate into 45 VOFP points.
The system is based on real success of fantasy players over the last four years of the NFBC. I'd love to do another article showing how well you'd do in previous years of the NFBC if you got specific numbers at the end of the year and how your point values dictate in which place you'd finish. I've done it in the past and it works out very accurately. Of course, an individual player's projected VOFP will only be as good as the projections it's based off of, but PECOTA is a good way to go, I think.
I guess you compare this to traditional dollar valuations, rejecting the latter on the grounds that they seem arbitrary to most fantasy players. But what about the ones who do play auction leagues? Does VOFP work better in auction formats (using some transformation)?
What am I missing?
Additionally, don't you miss the idea that each category has a barrier to start scoring points? For example, with saves, last place will have just ten and most pitchers get your zero. With runs, last place might have 800, but getting to 600 might take zero effort at all. Once you decide to compete in a certain category it's going to take varying amounts of VOFP just to compete. Then each stat BEYOND that barrier number helps you.
Yes, there are thresholds for what you want from each category. The goal is to use VOFPs to build a balanced team that is at or above zero in every category. But you should never be far above zero in one category while far below in another because you can only go so high before there are no more points to earn while there is plenty of room to fall. Part of the purpose of the stat is to see if you are misallocating resources.
That's kind of why positional scarcity can be important.
However, the basic idea of trying to figure out what fantasy stats contribute the most to victory is very interesting, and so the article overall is a plus for me.
Oddly, I think that this article would have benefited from more tables to explain some of the examples like Cabrera.
Overall I find this to be rather difficult to follow. I think that it's a valuable article, but the writing needs some work. Undecided on if I'll give it a vote or not.
The idea behind VOFP intrigued many fantasy players when Alex Patton first introduced the idea more than two decades ago with "Standings Gain Points", which measured the number of projected standings points that a team should gain by rostering a particular player.
In 1997, Art McGee further refined the concept with Marginal SGPs, which incorporated superior auction pricing into the calculation. McGee's How to Value Players for Rotisserie Baseball nicely outlines the required method to derive values for a particular league.
Gene McCaffrey and Justin Eleff of "Wise Guy Baseball" then applied the SGP idea to the Diamond Challenge game offered by Fanball.com (neé CDM) in 2000, demonstrating how to develop a system for ranking players over multiple categories for a national contest. At best, VOFP merely translates these principles to NFBC.
The utility of VOFP/SGP also decreases in direct proportion to league size. You can estimate baseline stats with some certainty due to the large number of participants in Diamond Challenge and the NFBC. Unfortunately, attempting such a calculation makes little sense in highly volatile online leagues with significant owner turnover and virtually none at all in keeper leagues, where rebuilding teams drastically skew category minimums.
As a previous comment noted, BP's Player Forecast Manager already thoroughly trod this territory through the application of general replacement level theory to fantasy value projections (a field largely pioneered by John Benson in 1993's Rotisserie Baseball: Playing for Blood).
Rather than rehashing a decades-old idea, Oakchunas could have provided more useful advice for the vast majority of fantasy owners by simply copying "Know your league's rules!" four hundred times in an assortment of fun fonts and formats for the whole family to enjoy.
I guess part of my problem is the list of players in a typical 2009 draft. I was surprised there was no Alex Rodriguez in there (since news of his injury came late in the typical fantasy draft season), and no Ichiro or Carl Crawford-types past Jose Reyes. I feel this article would've been great if it had discussed how much Ichiro's batting average helps a fantasy team more because he walks so rarely that his batting average carries additional weight. I also wondered a bit about how useful a VOFP system would be in a 390 team league when the difference of a home run here or an RBI there could mean the difference between 40th and 60th place. It seems that a system like VOFP would need to be more precise in such a large league, and not less precise. I also thought some kind of statement should've been made that a home run also leads to a run scored and an RBI generated, so should garner some additional weight/consideration/calculation.
I did like the writing style overall and I thought this article was much better than some of the other submissions... but I also felt a bit more could've been done. Still, it'll get my thumbs up.
By the way, am I the only one that prefers my acronyms pronounceable. I can say "vorp" and "pecota". I can't say "vofp".
Without those details, the article was a bit like when they use swap-ins in a cooking show: here are all your ingredients, this is a general description of what we're cooking...then boom! the finished fantasy ranking system comes out of the oven, pre-made. I would've liked to see Brian cook.
As for the calculation, I look at how many homers there are between 40th and 350th place each year and do some simple division to see how much it moves you. When you graph the data, you see a steady progression, except at the ends. In other words there isn't a larger number of homeruns between 50th and 100th place than between 200 and 250th place so a hard number, like 4.08 is relevant here.
However, as someone who started doing auction league this season, there's no going back. I'm not sure how this would apply to that style and the strategies used there, where as PFM does that specifically for me with that program.
Well written, but the the topic pulls it down to a no vote.
VOFP and drafting strategy is a whole other article or series of articles. I was more interested in just introducing the conceppt, given my limited space here.
But he tried to do something fairly unique and analytical and in line with the week's topic.
For me at least, what Brian did is "new", and I thought it was both well-done, and well-written.
Brian’s method is easier to use than Tim's, but Kevin, Sky, and others have found some major holes. Brian admits he needs to make spreadsheet adjustments during the draft.
I didn't mind the "woot, woot", but overall in the head to head fantasy tool department, Tim's article was more fun to read.
I agree that auction dollars are fairly worthless when it comes to snake drafts, however.
I agree with the judges that Ichiro!s of the world deserve a slight boost for the obscene number of ABs that they contribute to BA. (Basketball fantasy already accounts for this in stats like FG and FT percentages by looking at FTA/game or FGA/game.)
This goes on the maybe pile.