A few days ago, I asked the question What is Wrong With Fantasy Baseball? The response from many was that nothing is wrong (if you think that, you’re wrong). In this piece I’ll take a stab at something that I hope there will be more unanimity with – that expected stats are ruining fantasy baseball.
Scene: Twitter
Tristan Cockcroft of ESPN writes about a potential breakout from Byron Buxton. One of the best analysts out there, Tristan spends 900 some words laying out his case as to why you should trade for Buxton.
I found the article interesting, and re-tweeted it.
The first comment to the retweet wasn’t positive. Hey, it’s Twitter and negativity rules the day, so I wasn’t surprised. If the negative reply was something like ‘sorry, I just can’t buy in after being burned so many times’ or ‘he’s never going to live up to his top prospect status’ or ‘I still have concerns with his overall approach’ or ‘he hurts himself too much throwing his body around the field’ I would have thought – valid point. I would have disagreed, but they would have all been responses that I would be able to understand. But the response didn’t speak to any of that.
The response was short and sweet: .284 xwOBA
My response: I’ll take 900 words, and actual analysis, from TC over an expected number every day of the week.
The response of the initial responder: oh lord
Analysis no longer matters because we have expected stats.
Welcome to fantasy analysis in 2019 folks.
The problems we face are legion.
A certain segment of people don’t like to read, so the idea of spending five minutes to peruse an article is anathema to some. Folks just don’t have time to devote to in-depth study. That’s not to suggest that everyone is lazy. Some people can’t be bothered, but a lot of us are just so damn busy – work, kids, life – that we don’t have enough free time to do what we love (play fantasy sports) so we search for shortcuts. Its understandable.
In most areas of life, people search for simplicity. I always thought if there was an afterlife that I would love to go to the library. Yes, I would read in death which tells you there might be something wrong with me. Why a library? Because in my vision there would be one book – the truth – on everything. I wouldn’t have to read 75 books to find out who killed JFK or 25 books on what actually happens in Area 51. There would be one book, and it would tell you the whole truth about whatever topic it was dealing with.
We have the same goal/idea in sports. We want one measure, one number, that puts everything on the same continuum. In baseball, we want to know who the best player is. We don’t care if one guy is fast, if one guy has power, if one guy is better with the glove or if that guy is a pitcher. All we want to know who is the best. This is why analysts have for years tried to come up with one, all-encompassing measure to put all the players into one rankings system (we’ve had attempts like Win Shares, Linear Weights and WAR to name but a few). Simplicity can be sexy (think a long black dress on a lady).
We have more data at our disposal than at any time in human history. I mean, you can just go to Wikipedia and inerrantly learn everything that has ever happened (slight hyperbole noted). The cavalcade of data is fantastic for those of us that have the time and desire to dig into it, but as I’ve noted, not everyone has the time to do that. For the majority of people, it’s all just white noise that is utterly confusing. Do you listen to A or B or C? Hell, maybe D is better in this instance, but next week E will be more applicable. I mean, how does anyone know unless they are a specialist in the field?
Given all that, it only makes sense that the consumer of a fantasy baseball product is in search of a reductionist position – one number that cuts through all the noise. Into that void has stepped the expected stats boon.
Let’s go back to the example above with Buxton.
How many of you know what xwOBA is? If I stopped you on the street and asked you, would you know what the abbreviation stands for? Even if you knew what it stood for, would you know what it actually is referring to? Would you have any idea how to actually calculate the number, or what goes into it? This is not me – Mr. Know Everything – putting you down. I’m merely stating an ever-growing concern which is everywhere in the world in 2019.
With all the knowledge we have at our fingertips we have access to data that we could never have hoped to have had in the past. We have data that we cannot possibly interpret properly without the background information, and frankly, very few have this background data. Think of it like this. Would you be able to do algebra if you didn’t learn rudimentary mathematics? We all fancy ourselves as handy, and I’m sure most of us could put up a towel rack in our house, but could you take a wall down to open up a room? Of course not because we are smart enough to understand that our roof might collapse if we removed a load-bearing wall. That’s why you hire a licensed contractor who knows how to do things properly. We should apply this same line of thought to our fantasy baseball analysis. Make sure you have the foundation set before jumping to the advanced stages. I’ll get back to this line of thought in a moment.
Before moving on… xwOBA means Expected Weighted On-Base Average. So in order to be tossing that number out there, as in the case of Buxton above, you need to first understand what wOBA, or Weighted on Base Average, means before taking the secondary step of turning to xwOBA.
Speaking of expected stats directly, they are everywhere. Thanks to the excellent resource that is Baseball Savant, we average folk now have access we could only dream about twenty years ago. We have Expected Batting Average, Expected Slugging Percentage, Expected wOBA and more. There no longer is a need to break down the components of a hitter because we can just skip to the conclusion and get what the number should be. It doesn’t matter what the number is in the real world because if we know the expected number, then surely we know what it should be and that means it’s easy to predict what the future holds. It’s that simple, right?
Don’t just think that users/players lean on the expected numbers either as there are a bunch of lazy ass analysts that are doing it too (or perhaps they aren’t lazy, perhaps they just aren’t very good analysts). This is my true gripe here, 1,200 words into the piece (I hope you are still reading). I believe that too many people – players and analysts – skip to the final chapter, quote the expected stats, and don’t bother to read the rest of the book. I’ve written about expected stats myself many times this season, but before you call me a hypocrite, let’s look back to the first time I wrote about expected stats this season, in which I included the following statement.
My cohort, Jeff Mans, is fond of saying something along the lines of ‘just look at the expected stats and you will find the players that folks are pushing in fantasy.’ He’s not wrong. While the measurements can certainly be used to help us paint a better picture of what is going on with a player, you shouldn’t take it too far, to the point where the overarching read you get on a player is that ‘well, his expected stats say…’ Piece of the puzzle it is folks, but it’s not the whole puzzle, so be careful not to treat it as such.
I’m all about instant communication. It’s why I pay so close attention to Twitter and our Chatroom. I want to reach you with the info you need, and to do so in a timely manner, but perhaps it’s the way we are conversing as player and provider these days that is causing the current issue? Twitter, while a great place for immediate info, has obvious limitations. With only 280 characters there is only so much you can share, so doing something “short-hand,” like using expected stats, I guess I understand (at least a little bit). At the same time, as an analyst, it’s cheap, easy and banal to lean solely on expected stats. You should expect more from your analysts.
HOW TO USE EXPECTED STATS PROPERLY
Taking a look around the fantasy universe, it’s clear that some folks lean on expected stats. Like, their whole point about a player being better/worse in the coming weeks is built upon expected stats. Look at the add lists for some folks each week. They will be populated with expected stat potential risers. At the same time, those players that the analyst is down on – wanna bet they are on the wrong side of the expected stats list?
Let’s take a look at how expected stats should be used by going back to the case of Buxton. Here is what a paragraph about Buxton should look like by incorporating expected stats into the mix.
Buxton is a tremendous talent, always had been. He’s had trouble staying healthy, and consistency has never been his thing, but it certainly looks like he’s taping into something in 2019. Buxton has a 91.7 mph exit velocity that is up six full mph from his career levels which is impressive. He’s also moved the hard-hit rate from 32 to nearly 45 percent as he’s flat out hitting the ball much harder than in years past. He’s also increased his launch angle from his traditional 12-13 percent range all the way up to 20 percent this season. It’s therefore unsurprising to see that Buxton’s expected slugging percentage is a career best .408, some .058 points above his career mark. Additionally, Buxton has increase his walk rate while significant cutting down his strikeout rate allowing him to post a 0.30 BB/K ratio which, while still a poor number relative to the league average of 0.39, represents a 33 percent improvement over his career level. He still has some steps to take, his .305 expected wOBA is below the .318 league average, but his increased rate of hard-hit baseball is certainly encouraging.
That’s analysis, with a sprinkling of expected statistical information. That’s how expected data should be used. If an analyst is doing the above, then they are doing a solid job providing you the context that should be giving with a player. However, some analysts seem to have lessoned the role of context believing that the expected number supersedes that data – it doesn’t. There is fantasy analysis everywhere these days. Some of the work being done today is astounding, a fact that I want to make sure I note. At the same time some analysis is at best sloppy/lazy or at worst simply uninformed. I would encourage all of you to be more discerning readers.
Expected stats are not the end game of analysis, not matter what some folks are trying to tell you. Expected stats have their place in analysis. They are certainly useful tools to ponder when trying to determine who a player is or isn’t, but unless you see actual analysis accompanying those notations of expected stats then it’s best to realize you’re being sold something that may not have lasting importance.
