The xFiles: “It’s A Skill Game, Jo”

“It’s a skill game, Jo.”

To quote the great Mike McD from the movie Rounders, pleading with his girlfriend that he’s not just a degenerate gambler carelessly throwing his money away playing poker:

“Why do you think the same five guys make it to the final table of the World Series of Poker every year? What, are they the luckiest guys in Las Vegas? It’s a skill game, Jo.”

It’s a skill game Jo pic.twitter.com/JSl9VjBcZg

— Collin Capone (@PokerCapone) September 20, 2022

The same is true for fantasy baseball. I’m in two large dynasty leagues, both having drafted several years apart with largely the same group of managers, and almost to a person the same cast of characters are in contention in both leagues. Are they just luckier than the others? Of course not. It’s a skill game, Jo.

There are different approaches to success

How great managers consistently ascend to the tops of leagues can vary.

Some managers are legitimately lucky. Hey, luck wins leagues and money too. I am convinced there are some managers who win in spite of poor process.

Many managers do a great job of listening to fantasy experts and using player raters, rankings, and projection systems on websites like this one. Please keep doing that. We’re here to help!

Then there’s now an increasing number of managers who make good use of modern day statistics to predict breakouts and bounce backs that help them add value to their roster and win leagues.

This series, titled The xFiles, focuses on the latter approach and the goals are as follows:

1. Increase awareness and understanding around increasingly popular advanced statistics.
2. Help you more easily identify and target players on the cusp of breaking out or bouncing back.

There’s no better time to up-skill than now

Given that most of us are rapidly approaching trade deadlines, now seems like as good of a time as ever to canvass rosters and waiver wires in hopes of targeting and adding an impact player or two with the ability to push our teams over the top and across the finish line.

With multiple deadlines on the horizon in my leagues, I decided to do a deep-dive this past week. Here’s what I found…

Good and even great players slump deep into the regular season

Really good players can slump and underperform deep into the season. This time last year, the following players were still mired in slumps or underperforming. I’ll use OPS here as a simple, single measure of their output at the time but will replace it later in the column with better, more advanced stats.

• Julio Rodriguez – .742 OPS
• Trea Turner – .681 OPS (!)
• Bobby Witt – .727 OPS
• Kyle Schwarber – .733 OPS
• Seiya Suzuki – .730 OPS

Each of those players had two things in common going into August and all of our trade deadlines:

1. They were largely rostered by frustrated managers who were willing and ready to pull the cord at a moment’s notice.
2. They had underlying statistics that pointed to better days.

Many of us remember how those players finished up the final few weeks of the season because we all either experienced the joy of their outbreaks on our own rosters or the misery of having to face them on others.

Just in case you forgot, here’s what those same players did from July 27 until the end of the season:

• Julio Rodriguez – .948 OPS
• Trea Turner – .961 OPS
• Bobby Witt – .967 OPS
• Kyle Schwarber – .967 OPS
• Seiya Suzuki – 1.001 OPS (!)

Fortunes can turn quickly, but you don’t have to be lucky to strike it rich!

(Before we move on, I want to quickly acknowledge that clearly there’s more to fantasy value than pure hitting ability, so while this column will not focus on counting stats like runs, RBI, or steals, it’s very rare that a player can hold significant fantasy value without the ability to hit. So, for me, I target players with hitting ability first and foremost. You can’t add fantasy value if you aren’t hitting or on the base paths.) 

Rebounds and breakouts are relatively predictable

So, what were some of the underlying statistics managers could’ve used to predict these players would turn it around?

Well, for one, it doesn’t take a genius to predict great players will snap out of slumps. But not all these players were considered great at this point last season. We all hoped Seiya Suzuki would be really good, but he hadn’t quite shown it yet.

Or had he?

If you had taken the time to look under the hood this time last season you would’ve seen Seiya Suzuki (along with the others listed above, plus many more) was already doing some really important things despite the lack of results:

1. He was demonstrating a disciplined hitting approach
   1. He had a low O-Swing rate (chasing pitches outside of the zone below 30% of the time), above average BB% (walking near or above 10% of the time), and a reasonable K% (striking out in the low 20% or lower range)
2. When he swung the bat he didn’t miss much
   1. He had a solid Swing-Strike % (below 10%)
3. He was making consistently hard contact
   1. He had a strong Exit Velocity (over 90 mph) and barrel rates (over 10%)

These aren’t hard and fast rules, but hitters who do those three things well are generally good. If they aren’t producing despite those indicators, you can safely assume the production will come with time, as it did with Seiya Suzuki. Again, not in every single case, but most. Not immediately, but over time. The data backs this up, not just with Seiya, but with most hitters. 

Two statistical indicators to rule them all?

But there’s an easier, quicker way to potentially sniff out a slump or a breakout candidate than digging through swing and batted ball rates player by player. Two very similar stats that take a lot of those three key things good hitters do that I listed above into account. Those two stats are wOBA and xwOBA.

Many of you are already familiar with these stats, but in case you aren’t, here’s a brief explanation:

Baseball fans embraced OPS a long time ago as a single stat measure of a hitter’s production and ability. But OPS, while useful, is kind of flawed. It does a poor job of accurately weighing OBP vs slugging, for one, which can make it a bit misleading when evaluating and comparing hitters. More importantly, for our purposes here, it’s not really predictive. It’s a measure of what a player has done, not what they are expected to do based on underlying data.

So we need a couple stats that help us better understand a player’s offensive impact AND one that is predictive.

Enter wOBA and xwOBA.

Enter wOBA and xwOBA.

wOBA (weighted On Base Average) is essentially a weighted on base percentage but it’s not just about how often you get on base. That’s OBP. It’s a more complete and accurate measure of a player’s impact per plate appearance, accounting more precisely for the impact of different offensive outcomes like walks, HBP, singles, doubles, triples, and HR.

For context, a player with a .310 wOBA is about league average. This season you’ll find players like Ha Seong Kim, Brice Turing, and Nico Hoerner all hovering around a .310 wOBA. Fortunately for the managers rostering those players, they add value with their legs because absent that they would all lose relevance in most fantasy leagues.

Players with wOBA in the .330 range are producing at solidly above average levels. They aren’t killing it, but they are having a positive impact on their teams and likely hitting at a level that makes them fantasy relevant.

Players with wOBAs in the mid to high .300s are having great seasons and producing at levels generally among the game’s best, especially as you get closer to .400. Bryce Harper has a .392 wOBA this season. Elly De La Cruz is at .367. Both are great hitters, but the gap there exists because EDLC simply makes less contact and that results in lower average and less impact than Bryce. Of course, in fantasy leagues he makes up for that with his legs, making him an equally relevant if not superior fantasy player.

Then, you have players who transcend the sport. True superstars. Hitters that regularly produce at or well above .400 wOBA. Aaron Judge currently has a .460 wOBA this season. His career wOBA is .415. To put that in historical perspective, Barry Bonds has a career .435 wOBA. Mike Trout .415. This should give you an idea of what a hitter is doing if they are posting a wOBA above .400. If you’re rostering Marcel Ozuna, for example, then you know exactly what this kind of impact looks and feels like. Ozuna has a .415 wOBA this season.

But the question is, can a hitter like Ozuna keep this up? Much like Seiya’s breakout last season, how can we predict whether he’s legitimately capable of this level of greatness or if he’s just lucky and destined to fall off and disappoint managers down the stretch?

This is where predictive stats matter. Stats that don’t just measure what a player has done, but what they should be doing and are likely to do going forward if all things stay the same and a player’s production normalizes over time.

But wOBA itself is not predictive. Again, it’s a measure of what a player has done, not what we can expect them to do.

A better stat for our purposes is xwOBA, (expected weighted On-Base Average) which takes the weighted real life production of wOBA (all the walks, HBP, singles, doubles, triples, and HR) and adds additional context that might account for a player doing good things (hitting the ball hard and running fast) but not seeing the results they probably should. So, xwOBA considers exit velocity, launch angle, other batted ball data, and sprint speed to add context to a hitter’s performance.

That means if, say for example, 2023 Seiya Suzuki is scorching the ball at good launch angles but only getting mediocre results, especially in the power department, xwOBA can inform us that he’s probably getting unlucky with balls in play and over time those things should level out, resulting in markedly better production. That was precisely what happened last season. 

Who is currently under-performing according to xwOBA?

Now to the part of the article that you probably care about the most. Which players are currently under-performing according to xwOBA? Who should you target in trades or in some cases on the waiver wire?

To determine who is under-performing their expected stats, simply lineup wOBA (actually production) and xwOBA (expected production) and look for gaps. This can be done to target over-performers too.

The list of players under-performing their xwOBA is long, but I decided to limit the list to players with at least 150 PA and a disparity at or above 0.025 percent:

• Ben Rice – .306 wOBA vs .386 xwOBA (0.080 disparity between wOBA and xwOBA)
• Alejandro Kirk – .270 wOBA vs .332 xwOBA (0.052)
• Christopher Morel – .299 wOBA vs .347 xwOBA (0.048)
• Tyler Soderstrom – .310 wOBA vs .356 xwOBA (0.046)
• Fernando Tatis Jr. – .358 wOBA vs .400 xwOBA (0.042)
• Taylor Ward – .309 wOBA vs .351 xwOBA (0.o42)
• Jesus Sanchez – .306 wOBA vs .347 xwOBA (0.041)
• Ivan Herrera – .318 wOBA vs .359 xwOBA (0.041)
• Dansby Swanson – .280 wOBA vs .320 xwOBA (0.040)
• Julio Rodriguez – .304 wOBA vs .341 xwOBA (0.037)
• MJ Melendez – .283 wOBA vs .320 xwOBA (0.037)
• Salvador Perez – .344 wOBA vs .378 xwOBA (0.034)
• Cal Raleigh – .310 wOBA vs .344 xwOBA (0.034)
• Andrew Vaughn – .292 wOBA vs .326 xwOBA (0.034)
• Jo Adell – .283 wOBA vs .317 xwOBA (0.034)
• Paul Goldschmidt – .293 wOBA vs .326 xwOBA (0.033)
• Corey Seager – .356 wOBA vs .387 xwOBA (0.031)
• Francisco Lindor – .352 wOBA vs .383 xwOBA (0.031)
• George Springer – .306 wOBA vs .337 xwOBA (0.031)
• Shea Langeliers – .305 wOBA vs .335 xwOBA (0.030)
• Austin Wells – .330 wOBA vs .360 xwOBA (0.030)
• Leody Taveras – .285 wOBA vs .315 xwOBA (0.030)
• Austin Riley – .334 wOBA vs .363 xwOBA (0.029)
• Lars Nootbaar – .320 wOBA vs .349 xwOBA (0.029)
• Josh Lowe – .278 wOBA vs .306 xwOBA (0.028)
• Andrew McCutchen – .319 wOBA vs .344 xwOBA (0.025)

You’ll notice some big names here like Julio Rodriguez, Fernando Tatis Jr., Corey Seager, Francisco Lindor, and Austin Riley. Tatis is out indefinitely and Lindor, Seager, and Riley have already started turning their seasons around. The opportunity to buy low there has probably passed. But there might still be some frustrated J-Rod owners. As the saying goes, “you miss 100% of the shots you don’t take.”

What I find more interesting are the non-elite players. That’s where the real opportunities exist for fantasy managers. I intend to do a deeper dive on some of the younger, less established players on this list in a future column, but there are some incredible buying opportunities up and down this list, especially those with bigger disparities. In shallow redraft and dynasty leagues, some of these players can be had with a simple add in free agency or a cheap player, prospect, or pick swap.

Again, I cannot stress enough that just because you see a player’s name is on this list, it does not guarantee they will bridge the existing gap between wOBA and xwOBA overnight or at all. Clearly, some of these players I would bet on more than others. Some have smaller sample sizes that could lead to ‘noisier’ data. Some have metrics that will fall off because pitchers will make adjustments to them and they won’t able to adjust back.

But, in this game, I look for every advantage I can find and xwOBA provides an objective measure of tangible skill, ability, and expected outcomes. You either have luck or you have tools and strategies that will set you apart. Remember, it’s a skill game, Jo.