Steve “Dr. A” Alexander bemoans the variety of fantasy basketball leagues, which makes ranking players in general basically impossible:
I’m in 11 hoops leagues this year, and every single one is different. There are Rotisserie, Head-to-head and Points-based leagues. There are 30-team leagues and there are eight-team leagues. There are five-category scoring leagues, and there are 14-category scoring leagues. Offensive rebounds, turnovers, 3-point percentage, triple-doubles and technical fouls are just some of the weird categories prevalent in many fantasy hoops leagues.
… Because of all of these factors and the fact so few leagues are identical, ranking NBA players in fantasy hoops is not a pleasant task. Dwight Howard is the prime example. In leagues that count free throws made instead of free throw percentage, and don’t count turnovers, he’s probably the No. 1 overall pick. In Roto leagues that count both FTP and turnovers, he’s virtually undraftable, unless you think you can win despite taking a guaranteed “1” in that free throw percentage. …
So the bottom line is that you have to take rankings as a rough guide, and not a bible when looking at whom to draft. It all depends on your league’s settings. Our Tiers are the easiest thing to keep updated accurately and are my favorite tool to use when drafting, but whatever rankings you use when drafting, take into consideration everything I mentioned above. Blake Griffin is a stud in points leagues, but is a bit of a Roto nightmare. And trying to come up with a general set of rankings in hoops is nearly impossible.
This problem is what RotoValue Analyst is designed to solve: customized valuations based on your scoring rules. The RotoValue pricing model takes into account what categories you use, as well as the roster sizes and your salary cap, to compute theoretical auction prices customized to your league’s settings. So in leagues where free throw percentage is a category, Dwight Howard ranks much lower than in ones where it isn’t used. In my review of 2011-12 top performers last month, I compared values for two 8 and 9 category leagues, but both counted FT%, so Dwight Howard wasn’t even the top center, let alone the top player overall.
How does RotoValue work?
The model starts by ranking each player in every category. If you’re the best in the league in the category, you get 1 point for that category, and if you’re the worst, you get 0. Counting stats are simple: whoever gets the most rebounds gets 1 point there, while if you don’t get a rebound at all, you’re at 0. Other players in between are simply scaled relative to the league leader.
Percentages are a bit trickier: you can’t simply take the highest and lowest values in the league – someone who goes 4-4 is not as good as someone who goes 180-200. So for those categories I consider the impact adding a player’s percentage would have to an “average” fantasy team. Thus Dwight Howard winds up getting 0 in FT%, because not only does he shoot poorly, he takes a lot of free throws, so his percentage hurts a team more than any other player. Last year, despite missing 12 games, Howard led the league in free throw attempts, so his 49.1% hurt a team much more than Ben Wallace’s 34.0%. As Dr. A. notes, owning Howard basically dooms you to finish last in FT%. I know from experience – in my 6-category league I owned him last year and in the previous season.
Jamal Crawford led the league in FT% at 92.7% among players with more than 200 FTA, but both Dirk Nowitzki (89.6%) and Kevin Durant (86.0%) were more valuable free throw shooters, because they took many more attempts.
Once I have the raw rankings in each category, I scale them to match the weights your league uses, and then I add up the category scores to come up with a total player value number. This value shows how much a given player helps a fantasy team. Of course some leagues are deeper than others, and roster composition also varies. What matters isn’t really the raw value of a given player, but his value relative to others, in particular the “replacement” level player at his position. At the end of an auction, somebody is left available for the minimum bid, and so what matters most is how much better a player is than that “replacement” player.
What the replacement level is depends on position, and, of course, the number of players on your roster. Overall centers (or catchers in baseball) may be less valuable than other players in raw terms, but if each team must start two players at each position, including center, then some starting centers will be less valuable than bench players at other positions. So the model ranks players in raw value order, and then subtracts the value the value of the replacement player at a given position from each player’s raw value to get a net value. In raw terms, Dwight Howard won’t be better than Kevin Durant or LeBron James (unless your only categories are rebounds and blocks), but if you need a lot of centers (and aren’t counting FT%), then he might be close, or even be worth more compared to the replacement level. And thus he could be worth a higher bid.
So once I have a net value of all players, I distribute the available total auction money in proportion to their net value to compute a RotoValue price. The league minimum bid has an impact: I assume that all bench players are worth the minimum, leaving a little more money to bid on better players.
If you’re not in an auction league, you can still find RotoValue prices quite useful, because a descending price order would in theory match a draft order (assuming the “best available” player is taken at each spot). And in addition to an ordinal ranking, which any cheat-sheet can provide, a price list also quantifies the differences in value between players. While I do compute and display prices to the penny, I don’t think it’s that precise – one can often make a good case that a player with a somewhat lower RotoValue actually had a better year than another with a higher one. And of course any pricing model is only as good as the statistics you give it. RotoValue lets you use current year, prior year, or (for RotoValue Analyst subscribers) preseason data, as well as its own projected statistics as inputs to the model. You can also capture stats over any particular date range (useful in seeing “hot” players recently, or comparing a player after he’s returned from injury). No model should be the final word on bidding or drafting decisions, but the prices RotoValue computes can help you inform your choices and decisions, not only at the draft, but even during the season.
Oh, and if you do wind up with Dwight Howard on your fantasy team, RotoValue Analyst lets you override the category weights for your league when computing prices. So you can set FT% (or free throws missed) to have 0 weight, and then recompute values. In a Rotisserie scoring style, once you’re last in a category, it doesn’t matter how far back you finish, so you might as well stockpile bad foul shooters who are good at something else. That’s how I won with Howard on my team 2 years ago!
RotoValue is offering full commissioner services for your league, including live updating stats, for an introductory price of just $50 for the full 2012-2013 season, no matter how many teams are in the league. RotoValue Analyst – which gives access to RotoValue prices customized to your rules, is just $10 per sport. So if you’re Dr. A., with 11 different leagues all with different scoring rules, one $10 fee could enable pricing for all the leagues hosted at RotoValue. Questions? Contact us at firstname.lastname@example.org, or visit the RotoValue Store!