I didn't want to do this, but when Mike Rathburn (who's been the best guide I've come across yet for DFS) didn't have a specific answer to my question today on SXM, I had no choice.

My question was how many points a player with an OPS of x is worth on average in FanDuel. Because one can take a player's platoon OPS, adjust for pitcher quality and park and come up with a rough estimate of who that player would be should he get that same matchup every game. Once we have everyone's situation-specific OPS or wOBA or whatever stat you want to use, we should be able to translate it into FanDuel points, adjust it for projected number of plate appearances, based on where he hits in the order (as well as the projected O/U run total for his team that day) and see how many points per FanDuel dollar he's slated to earn. From there it should be possible to pick players who provide the best values, given their costs.

To do this, I started with Steamer's rest of the season projections and translated those into Fan Duel points. For pitchers it was easier because Steamer projects games started, so I could divide the season totals by the number of starts. Here's how the pitchers project on a per-start basis (I deleted any pitchers who were projected to have relief appearances because it was impossible to calculate per-start numbers for them, and I also deleted any pitchers projected for fewer than 73 IP because most of them were scrubs:)

When you compare the expected pitching output to the cost, nearly every pitcher earns more than a point per $1000 spent. Considering you have a $35,000 budget and need roughly 35 points to cash in a 50/50 or Double Up, almost are pitchers are profitable, certainly when compared to hitters. But because everyone has to roster a pitcher, you need to compare pitchers to each other. Here are tonight's starters (the ones who qualify under the projections criteria) sorted by net profit, where profit is defined as expected points earned*1000 minus cost:

Keep in mind the expected points earned is not adjusted for opponent, park or opposing pitcher here. Also remember for a 50/50 or cash game, you don't necessarily need to max out profits, you just want to lock in some significant profit, i.e., low volatility is desirable. For that reason, Shields or Lackey might be better plays than either Anderson or Walker. In a big tournament, the lesser-owned, high-profit players make more sense.

Let's look at hitting though because that's where things get more interesting:

A few caveats here:

(1) Because the in-season Steamer projections don't include games played, I estimated games played by dividing plate appearances by 4.2. I used 4.2 because the average team last year had 6131 plate appearances, and 6131/(162*9) = 4.205. I divided total projected FanDuel points by projected games played to calculate projected per-game points.

(2) To be more precise, you'd want to adjust further by batting order slot. As of 2009 - and maybe this has changed since - the average number of plate appearances per order slot was as follows:

So when figuring out a No. 2 hitter's projected fantasy points, you'd want to multiply the per-game total by 1.07. For a No. 9, you'd want to multiply by .89. I'm fairly sure the Steamer projections account for this in cases like Trout and Stanton where their general lineup slot is known. But for players who move around or surprisingly find themselves hitting first on a given night, this adjustment should be made.

(3) You'd also want to adjust for team quality, i.e., teams in better offenses and in better parks generate more plate appearances. Again, I assume Steamer accounts for this, but if an average team were facing a bad pitcher, you'd expect it to see as many appearances as a top offense.

(4) I used only players projected for 200 or more PA.

Once we had all that, you'd obviously focus on platoon splits and park effects and calculate a player's rough projected OPS or wOBA for a given matchup. But setting that aside for a moment, let's return to the original question: How many FanDuel points does a particular OPS or wOBA portend? Stolen bases complicate this somewhat, but if we compare two low-steals options, say Jose Abreu (projected .395 wOBA, 3.4 FD PPG) and James Loney (.312 wOBA, 2.07 PPG), we can start to map out roughly how much we can expect at each real-life production level.

In Monday's game, Abreu costs $4100, while Loney will run you $2,700. Both players project as losing investments - as do essentially all hitters - but Abreu earns roughly $3,400 (3.4 *1000) for a net of -$700 while Loney earns $2,070 for a net of -$630. Straight up, they're largely a wash, but on a percentage basis, Loney is far worse, so you might be better off spending more of your hitting budget on Abreu and minimize your overall loss. Keep in mind that's before we add in splits.

As such, Loney isn't really a .312 wOBA player because he's much better against righties, and he plays half his games in a pitcher's park. So if he's facing a weak righty, say at Arlington, maybe he's a .330 wOBA player. And Abreu, at home against lefties, might actually be over .400. Moreover, Loney's average points on FanDuel aren't exactly the same as that of every .312 wOBA hitter without speed, so we'd be better off taking an average of what OPS or wOBA equals what amount of FD production. But my thesis is once we narrow down who we think a player is given a specific matchup wOBA/OPS-wise, we should be able to assign him a generic amount of FD points for that game and compare that to his salary. In fact, this is what the RotoWire Lineup Optimizer largely does (though I doubt it uses OPS or wOBA per se, and it's processes are opaque to users.)

Once the DFS player gets comfortable translating OPS or wOBA or whatever his metric of choice is, e.g., I know McCutchen at home against lefties is an 1.100 OPS player, and assuming I find 1.100 OPS players score 4.5 FD points on average, before adjusting for his speed, then it's easier to compare him value-wise to other options. Eventually, one should get a feel for the sweet spots, e.g., a $3,300 player with an expected OPS of .930 hitting second or a $2,200 one with an expected OPS of .840. hitting sixth. Perhaps frequent players already locate these sweet spots instinctively, but you're more likely to construct optimal lineups if you have a decent map.

When I have more time, I'll try to identify and post about specific sweet spots and look more deeply into how to structure rosters as a result.