The following is a guest post from Matt Strachman, a man who plays in more fantasy leagues every year than most of us have in our entire lives. This gives him a unique perspective on fantasy management, which he has agreed to share with us in a six-part miniseries. This is Part 1.
Baseball Professor’s own Bryan Curley reached out to the fantasy community for access to their league data, and after I sent over a file with information from 22 leagues, he asked if I would be interested in sharing some thoughts on my experience. Obviously fantasy baseball team management is a very complex concept, but I’ll attempt to shed some light on as many strategies/scenarios as possible while also discussing the best approach to playing in so many leagues.
Half fantasy baseball degenerate, half budding sabermetrics virtuoso, the name is Matt Strachman, however most people know me as NextLevelFantasy on Reddit or @HisDudenessOfNY. I graduated from Cornell University with a degree in Industrial and Labor Relations, yada yada yada, and decided to immerse myself in the industry and statistical data analysis.
Long story short, I used to pump out fantasy content, and have written here and there for about six years. The pen has dried up recently as my time is now spent managing an ungodly number of teams and crafting both yearly and daily baseball projection algorithms. I also founded the official multi-division reddit fantasy baseball, basketball, and hockey leagues with English Premier League style relegation from year to year. So to conclude this half-assed intro, the current tally for 2013 is as follows:
- 22 baseball leagues (while commishing another 27)
- 44 football leagues
- 31 basketball leagues (while commishing another 44)
- 5 hockey leagues (while commishing another 12)
Whew. I also organized a World Baseball Classic snake draft and grind it out on the daily circuit, just to cement the point I am entirely too obsessed with fantasy sports. By no means do I view myself as an accomplished man transcribing the details of a shiny body of work. I’m still just a snot-nosed 20-something on the base of the mountain, but my goal is certainly peak or bust.
Complete Resume of Experience
- 8, 10, 12, 14, 16, 18, 20, 30 team leagues
- Auction, Snake, Linear
- H2H categories, H2H points, Roto
- Redraft, Keeper, Dynasty, Ottoneu, Salary cap
- Standard 5×5 cats, OBP, SLG, OPS, Hitter Ks, Net Steals, Fielding %, Game-winning RBI, Grand Slams, Hitting for the Cycle, K/9, K/BB, Net Saves, Quality Starts, Losses, Holds
Using Projection Systems to Value Players
Gathering player projections is a great first step, but knowing exactly what those projections mean and accurately applying them to your unique league format is even better
Having a well-rounded understanding of players and teams is step one, but league settings must dictate your actions. While I am a big fan of wOBA, FIP and all those beautiful metrics that give a clearer picture of overall production, a player’s value is only relative to the stats in play. The size of the league, and in turn the size of the roster/bench, categories, moves limits, innings limits, etc., spell out the ideal team creation and management techniques. Most research and rankings refer to standard categories and you should customize your own rankings if you go beyond 5×5.
Proper pre-draft analysis deserves a novel on its own, but you should utilize multiple rankings and cross reference them with various projection systems like PECOTA, ZiPS, and Steamer. Depending on the platform you are using, tracking the disparity between their site rankings and overall ADP is a must. Read up on player analysis, since scouting adds a human element to the numbers as well.
A projection or ranking on its own is helpful, but the confidence scores placed on them and understanding why they are projected to do something is vital. Performance volatility, regression in which direction (for better or for worse), safety or high risk/high reward, etc. are somewhat quantifiable pieces of data that can help you make more informed and customized decisions during and after the draft. There is a ton of information out there, but besides the wonderful analysis found at Baseball Professor, I really enjoy the content at Baseball Prospectus and Fangraphs.
Absorbing news and expert analysis is great, but you should understand how they came up with those opinions. Work on familiarizing yourself with the statistics directly as opposed to relying purely on expert analysis. Research the building blocks of baseball and what statistics actually indicate. There are a ton of sabermetrics stats, the point isn’t figuring out which one is the best but instead what they each say. The fangraphs library is a great glossary to refer back to.
The Book by Tom Tango is a bit dry but a must read for serious sabermetricians. The Historical Baseball Abstract by Bill James, Baseball Between the Numbers by Baseball Prospectus, The Baseball Economist by JC Bradbury, and Curveball by Jim Albert and Jay Bennett are also chock full of knowledge. Keep notes and create an excel file with multiple sheets to keep track of your thoughts, revelations and relevant links.
There are a number of ways to value a player for your league including roto points above average, value above replacement, standard deviations using Z-scores, FIC scores, Point Share, WERTH, SGP, PAY, etc. My player valuation is slowly becoming more mathematical with the development of a data crunching program. While it wouldn’t be too difficult to calculate, grasping the methods while half-assing manual estimates using the different methods is better than nothing. Value above replacement is probably the simplest system to roll with on the fly during a draft, especially snake drafts where you can consider the needs of managers around you.
One random thought on Z-scores: You should more heavily weight the scores for the more predictive statistics.
Taking a step back, though, the more fantasy baseball you play, the more intuitive you become. The valuation systems are important, but they will only get you as far as the accuracy of the projections and assumptions. You should understand what “average” is in a standard roto league and use that as a baseline to ensure you don’t get too top-heavy with certain stats. At the same time, you can convert value into the stats you desire via trade so it isn’t necessary to get too obsessive sculpting your lineup.
Five more installment’s of Matt’s experienced-based strategy will be forthcoming, so keep checking back and make sure to comment with your strategies below!