Monday, April 04, 2011
Should I make this trade?Posted by Brad Johnson at 5:12am
With the real season underway, it is time for fantasy advice to turn from the ever-popular topic of the draft to in-season strategic moves. One question I see posed to fantasy gurus all the time is, "Should I make this trade?"
Frankly, and with the exception of challenge trades, you are in a much better position to decide that than even the most expert of experts. You know your team, your opponent's rosters, how both you and they will behave during the season, and various other insider advantages.
As such, today we will focus on a simple thought process to help decide whether a trade proposal is right for you. This guide is designed for simple roto leagues, although the same methods can be adjusted to any league format.
First and foremost, if you are in a serious league, I highly recommend conducting trade negotiations via email or chat rather than the propose and counter-offer tools offered by the main fantasy sites. In my personal experience, this speeds up negotiations by allowing both parties to propose offers before running the numbers.
The first step to making a trade is a basic smell test: Is this offer enough? Alternately stated, can I get more for this/these player(s)? We all know people who stop at this step, the kind of fantasy players who will take virtually any offer that smells fair to them. Don't be that guy!
This step is independent of actual statistical analysis and focuses on how you believe others in your league will assign name value to a player. For example, I am in a 5x5 OPS league. Under such conditions, Ichiro is actually expected to be slightly worse than Rajai Davis, yet it's very likely that name value could result in an owner getting twice the return for Ichiro. Obviously, the more "expert" the league, the less this comes in to play.
The second step is a very basic analysis using your personal replacement level. It is quite likely everyone reading this article already employs some form of analysis similar to this. The easiest way to explain this is to illustrate it with an example.
Let's say you own Adrian Gonzalez at first base, Casey McGehee at third base, and Adam LaRoche is on your bench. Now let's say somebody offers you Evan Longoria for Gonzalez. This trade offer passes the smell test, since both players were taken at similar picks and costs in most drafts. Next is to test if this helps your team. We can do this with a basic equation:
(Longoria - McGehee) + (LaRoche - Gonzalez) = X
If X > 0, the offer may benefit your team.
If X = 0, the offer appears neutral.
If X < 0, the offer may harm your team.
In practice, you would run that equation for each stat your league considers. Using Oliver projections and a standard 5x5 format, the results of that are -1 run, -3 home runs, -6 RBI, +8 stolen bases, and -.035 batting average (not weighted by expected at-bats). Those results are inconclusive because the third step is to evaluate how the expected changes affect your roto totals.
However, before we move along to step three, there is one important thing to note. Leaning on a single expected stats line is the most simplistic means to using this form of analysis. Some of you may prefer to add sophistication to the model by introducing confidence intervals. I like to set a range of stats that I expect the player to achieve with 50 percent confidence*.
For example, Oliver expects Longoria to score 85 runs. I might think there is a 50 percent chance he will score 75-95 runs. The distribution does not need to be evenly spread around the median, either. Longoria is projected for eight steals, but the range I would set is seven to 12.
*There are other methods to employ a range of expected values that are more statistically robust, but I find this gets the job done for my personal needs. Feel free to leave your own more robust methods in the comments.
Let's move along to step three now, evaluating your roto categories. In our previous example with Longoria and Gonzalez, we expected to gain eight steals while losing a run, three home runs, six RBI and about 35 points of batting average (remember that average is a rate stat, so this does not reflect the actual team-wide loss in the statistic). It is possible to dream up scenarios where taking a trade with these outcomes is a positive, negative, or neutral decision.
For example, if a team is strong on paper in runs, home runs, and RBI but lacking in steals, chasing Longoria's upside on the basepaths may be useful. If a team already has Jacoby Ellsbury and Jose Reyes, the other categories are probably more important.
Like with step two, there are various layers of sophistication that can be added. You can consider anywhere between the simplistic perspective of, "I need more stolen bases but have enough home runs" to the time-consuming method of projecting out every owner's expected stats totals. Things can be made impossibly complex by adding confidence intervals like in step two. Personally, the leagues I play in aren't for high enough stakes to go far beyond a simplistic approach.
This brings me to step four, a step that in my experience is criminally underused: Evaluating what your trade does for your opponent. Basically, we are just rinsing and repeating with steps two and three from your opponent's perspective. Sometimes this analysis is unnecessary. Perhaps you are in second place and your trading partner is hopelessly behind in eleventh. It can even be beneficial to help buried teams if it could take points away from your principal opponents. Just try not to get yourself caught in the cross fire.
At this early point in the fantasy season, even teams that are bad on paper could turn into contenders, so it is probably best to do your due diligence. Returning to our previous example, let's say the owner trading Longoria also has David Wright while his best first baseman is Lyle Overbay (yes, unlikely). In such a circumstance, you would be giving your competitor a huge boost in four categories, something you probably want to avoid doing.
Circumstances like these are fairly common too, if less exaggerated. I always charge a significant premium when "helping" an opponent, usually by asking for 120 percent of my sunniest expectation for a player.
As an example, my sunny projection for Ben Zobrist this year is 95/22/95/25/.285. I would ask a desperate trading partner to pay for something more like a 110/27/110/30/.305 player. This almost always prompts them to walk away, but unless you have your own dire need, this is not a bad thing. If one of your main competitors wants to trade with you, take advantage of the opportunity to make yourself better at their expense.
In practice, the gains your opponent realize from a trade are likely to be significantly smaller than Gonzalez over Overbay. However, it is still important to remember the ultimate goal of the season, winning. Try not to help an opponent beat you when making a trade. If a trade helps your team but also helps a main competitor even more, you should probably reject the offer.
The purpose of this brief guide is to discuss some methods for identifying when, and when not, to make a trade. One assumption of this guide that is not necessarily true is the ability to form a fairly concise expectation about a player's final roto stat line. Clearly this will not always be the case.
Furthermore, this is certainly not the be-all, end-all to trading strategy. In fact, you should insert your own techniques, proclivities, and idiosyncrasies into your personal trading strategy. After all, in fantasy baseball, the wisdom of the crowds can have a huge effect on the behavior of owners. If you are thinking outside the box, you may be able to leverage an unexpected advantage. Nevertheless, it is helpful to have a general process in place to help determine when a trade makes sense.
As always, please feel free to share your thoughts, questions, or strategies in the comments. Hell, feel free to ask for trade advice, too. Just be advised that you are better positioned to know the correct answer.
Follow Brad on Twitter @baseballAteam. Email him at pitchin432 AT Yahoo.com