Unicorn hunting - WAR

Editor's note: cdb is working on a series of Fanposts that looks not just at players' stats, but how often players achieve those stats in the last five years. The less often it has been achieved -- like Aaron Judge's 11.6 fWAR season in 2022 -- the higher it ranks. I thought it was a pretty neat way to show a truer "average" player. Looking at the histogram, you can see that highest number of players in the league in this five-year timeframe fall in the just under 2 fWAR range (though 2020 might have something to do with skewing this). 2 fWAR has long been considered the production of the average MLB player. They have already added to this analysis in this follow-up post and I look forward to reading the rest of it and thought you all might as well!

I also really love that they were inspired by ORSTLcardsfan. That is why I really try my best to share and promote these Fanposts -- we have such a talented community and we get such cool stuff to read! Please enjoy!

The recent amazing fanpost from ORSTLcardsfan has inspired me.

I need to force myself to brush up on some of the latest and greatest stats, and look at some data to become more familiar with non-cardinals players. I also want to understand 'good' and 'not', when i see numbers.

To do so, i am going to do a bit of playing around with fangraphs data. I am pulling data from the site using the R baseballr package. I can pretty easily download leaderboards for each year and merge them. Happy to post any R code if there is interest.

I am starting with the most recent five seasons of data. 2019 - 2023, batters only, with all batters with at least 200 plate appearances. This filter basically gives us all non-pitchers that are at least strong time share personnel.

I want to then look at the range of values for a given stat, and explore a bit who the outliers are - the unicorns, if you will. to do so, i will plot the data as a histogram, showing the frequency of each value range, broken into ~ 50 bins. i will also report the top 1% of all values so we can see who the unicorns are and in which season(s).

I am going to start boring: WAR.


this plot is a histogram, for those new to the format. each season is summarized by a single WAR value. The x-axis on the bottom represents the WAR value, and the y-axis how frequently that WAR value range was observed. We can see that there are a few reeaallly high WAR seasons, and we may want to know who they are. We can also see that do the left there are a relatively few seasons with negative WAR. Most of the time, those players don't play much - b but remember we have a 200 PA filter here, which gave them enough time to really stink up the joing.

We can also see that the vast majority of the player-seasons are between zero and one WAR. this pretty much defines 'replacement level' in the 'wins above replacement (WAR)' name. If this were not the case, WAR is being miscalculated or mislabelled.

I will report the extremes in table format:


And we can now see our very own Goldenado of 2022 - pair of MVP nominees! This is a trial. i want to see that i can do this. now that i have the data, i can do the same for essentially any fangraphs data type.

for fun, lets look at the reverse unicorns: OUCH!!


Jurickson Profar was a VEB crush for a while... Miggy Cabrera :-(

I'm not going to spend too much time talking about this. we are all pretty familiar with the best players on the planet. What I am interested in, now that i know how to do this, is to look at some of the more esoteric stats, and find some of those players that are really good at one or two things, and dive a bit deeper into those players. let me know if you have any requests for fangraph metrics to look at.