Elite Relievers in an Ottoneu Points Format

About 24 hours ago, plus or minus zero minutes, I posted on the topic of elite relievers and when such players should be purchased in a draft. I recommend skimming that first as some of the points are pertinent to this article.

That piece was designed to offer some general tools and thoughts for those in relatively standard leagues. However, FanGraphs offers a custom fantasy game called Ottoneu – perhaps you have heard of it. Ottoneu offers four scoring systems, two of which are scored by points. This article is about the value of elite relievers in those points formats.

I’m going to take a soft-science approach to this piece. I could put together some price sheets that place a dollar value on relievers and then compare it to position players, but I’m going to leave that up to the reader. For one, I’ve put together many price lists for points formats and have never found them to be directly usable. Sure they inform me of some things, for example Josmil Pinto might have been the best fantasy catcher last season if you used all of his starts plus a replacement level catcher for all other games. That’s nice trivia, but it’s not very actionable. If I have serious reservations about using my price lists, I’m sure as hell not going to share them with you.

If you are interested in building out your own list, I recommend some of Chad Young’s work on the topic. Specifically, here, and if you have FG+ access, here. The first post should give you enough fodder to build a relatively robust sheet with the help of Excel. The second spells things out in more detail.

This topic was brought to my attention by a reader, Trey Baughn. Apparently, there is a bit of an argument raging in his league – is it better to spend money on an elite reliever or an outfielder? Rather than do what’s probably my job and just answer that question definitively (I’m a contrarian, definitive doesn’t exist to me), I’m going to lay out a basic methodology and let you decide for yourself. Essentially, I’ll be the proverbial fishing teacher. It’s up to you to decide when and where to do it.

One final caveat. There are two types of points formats in Ottoneu, FanGraphs Points and SABR Points. I play FanGraphs Points so I’m going to use those values for examples. Pitchers can have wildly different values in the SABR format, so please keep that in mind.

From a high level, the decision of whether to employ one player or another, regardless of position, depends mostly on projection and replacement level. Obviously, other small factors are at play, but your range of projections, median/mean/mode expectation, and the position’s replacement level are of paramount importance.

Let’s start by dismissing the projections from this argument. Whether you use Oliver, ZiPS, PECOTA, Stanley, Steamer, or your own blend of secret sauce, no projection system is going to be perfect. Everyone uses some form of projection to pick a team. If you want to do a rigorous analysis, pick a system that allows you to express your expectations in points, innings pitched, and plate appearances. If you don’t want to take that step, go ahead and use your gut. That’s perfectly fine, people win leagues with gut drafting all the time. You can probably also click to whatever is next on your reading list (don’t actually do that).

So now that you’re committed to forming a projection, let’s talk about replacement level. Any points league is really all about Points Above Replacement (PAR). Just like with real baseball, there is a level of talent that is freely available for anyone to acquire at any time. In an informal survey of leagues, I found that team wide replacement level is about 12,000 points. Hypothetically, that should be a $40 team.  The target point total where you should feel pretty confident of a win is 19,000 points. So your goal is to spend $360 in a way that improves your PAR by 7,000 points.

Replacement level at each position is driven by the roster. For example, in Ottoneu, 12 teams have five active relievers. Most owners also hold zero to three relievers on the bench. So the replacement level reliever is generally around the 75th best.

A representative replacement level reliever last season was Robbie Ross. He was worth 351 points and 5.61 points per inning. The creme de la creme of relievers, namely Craig Kimbrel, Kenley Jansen, Aroldis Chapman, and Greg Holland, were worth 600 to 750 points and 9.5 to 11 points per inning. Theoretically, you can pay one dollar for Ross and a replacement level performance or $20 for Kimbrel and 371 PAR (Kimbrel was worth 722 pts last season, 722 minus 351 equals 371).

Let’s pretend that we’re back in the 2013 offseason and have perfect foresight. Kimbrel’s $20 price tag is a hair over five percent of your hypothetical free budget ($19/360). 371 points is also a hair over five percent of the 7,000 PAR we’re trying to buy. So on the face of it, Kimbrel was a break even investment last season. However, that performance was nearly as good as it could get. In reality, various risks should have reduced your projection of Kimbrel below what he actually contributed. Perhaps you should have been willing to spend $14 but not $20.

But… … …There’s actually another consideration for which we need to adjust. Even in a fresh Ottoneu league, some of the best players by PAR will also be among the cheapest. See Josh Donaldson last season. That functionally means that some teams have extra money available in their quest for 7,000 PAR. We call that inflation.

Trey shared a document with me that lists the common players among Ottoneu winners last season. The top of the list is dominated by a certain type of player, guys like Donaldson, Adam Wainwright, Shin-Soo Choo, Chris Davis, and Francisco Liriano. They were all very under-priced. A team that wins an Ottoneu league almost has to have one or more of those diamonds in the rough.

If you factor that into your planning, you might suddenly be willing to gamble on Kimbrel putting up a fourth straight $20 season. Basically, buy some of your PAR at cost and some via lottery tickets. Consider this, if  it was possible to buy all of your PAR at cost while factoring in risk, then that means the other owners in your league are extremely inefficient. I don’t doubt the presence of inefficiency in any Ottoneu league, but I have trouble accepting that the inefficiencies are that egregious.

Coming back to the question of relievers or outfielders, it’s purely a function of PAR per dollar. It doesn’t matter where you get your PAR as long as you get it. If you can spend $20 and get 371 points from Kimbrel or $20 on an outfielder and get 250 points, then the choice is clear. As the PAR per dollar merges closer, the decision becomes harder. At some point, the riskiness and high variance output of a reliever is going to cause you to pick the outfielder. The exact point depends on your projections, specifically the range, margin of error, risk factors, and other considerations. I don’t recommend actually being that detailed in your analysis, but those are all variables in the equation.

Your decision to roster an elite reliever in an Ottoneu points league comes down to PAR and dollars per PAR. That’s the decision criteria with any player. In reality, each owner will pay list price (or maybe more) for a number of guys like Kimbrel who promise elite performance. The owner who wins the league will probably also hire a number of players who substantially outperform their cost. That’s why we play the season, otherwise we could just make an OOTP league and rid ourselves of the daily grind of fantasy baseball.

Limitations: The 12,000 point team wide replacement level that I cite is based on last place finishes. It would have been more theoretically robust to calculate the replacement level of each position and factor in some small degree of subbing. But that is a lot more work for seemingly very little gain. Additionally, replacement levels can vary by season.





You can follow me on twitter @BaseballATeam

8 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Clayton
11 years ago

Last year Chad Young used this format for calculating values PAR * (4500/39680)

Here’s the article. http://www.fangraphs.com/plus/auction-values-for-all-three-ottoneu-formats/

For the life of me, I can’t figure out where that 39680 comes from. Can someone better at Math help me out here?

This year, using Steamer projections and that equation above, I have Kimbrel at $29. I doubt I’ll ever shell out that much for a reliever, I prefer hunting for Mujica and Uehara.

Clayton
11 years ago
Reply to  Clayton

I should clarify that this is in Fangraphs points leagues, and that calculation is (PAR*(4500/39680)). Sorry for the sloppiness.

Clayton
11 years ago
Reply to  Clayton

And I just realized that Brad linked to that article. Sorry, for the second time.

Homer Simpson
11 years ago
Reply to  Clayton

Get it together, Clayton.