Advanced MLB DFS Strategy: Pitchers Due For Regression

Thunder Dan Palyo identifies some of the most important stats to look at in order to figure out which pitchers to attack with hitters in MLB DFS an identifies some prime candidates for positive or negative regresssion.

The month of May is halfway over and most MLB teams have now played at least 40 games or 25% of their regular-season schedule. What that means to me is we are starting to get a larger and more reliable sample of data for the 2021 season.

This time of year is a great time to take a look at pitchers and look for some guys whose surface stats don't match up with their advanced stats. I am doing this primarily through an MLB DFS lens, but if you are a season-long player then you are constantly trying to re-evaluate players during the season and assess their future value as well.

This article is a follow-up to one that I wrote here earlier this spring for our MLB DFS strategy guide. If you ever want to chat with me about MLB, DFS, or fantasy sports in general, hit me up on Twitter @ThunderDanDFS.

• Fantasy baseball trade analyzer
• BvP matchups data (Batter vs. Pitcher)
• PvB matchups data (Pitcher vs. Batter)
• Who should I start? Fantasy baseball comparisons
• Daily MLB starting lineups
• Fantasy baseball closer depth charts
• Fantasy Baseball live scoreboard
• Fantasy baseball injury reports

Review of Stats Used

Stacking against a pitcher who is coming into their start with a 6.00 ERA and has allowed 10 homers in 8 starts is easy. It's almost too easy, meaning the opposing offense is going to be an obvious stack for tournaments (and perhaps still the right one). But some of those pitchers with bad stats may have positive regression coming, too, meaning a chalky stack against them might be a bad idea.

One of the best ways to differentiate your lineups is to find pitchers whose surface numbers don't look bad, but whose advanced statistics suggest that negative regression is coming. Here are some of the stats that I look at when trying to identify those pitchers on a daily basis.

1. ERA estimators - SIERA and xFIP: These are both better indicators of a pitcher's performance than ERA, which is an inherently flawed statistic. SIERA stands for "Skill-Independent Earned Run Average" and XFIP stands for "Expected Fielding Independent Pitching." Both metrics give a more comprehensive look into how a pitcher performed than just how many runs they allowed and factor in variables that ERA doesn't (such as luck, defense, park factors, etc...)
2. BABIP (Batting Average on Balls in Play): This stat can be used for hitters and pitchers, but can help in determining how lucky or unlucky a pitcher has been with how often they get outs on their pitches that are put into play by hitters.
3. WHIP (Walks + Hits Allowed per Inning Pitched): This lets us know how many baserunners a pitcher is allowing on a per-inning basis.
4. LOB% (Left On Base Percentage): This is often known as a "strand rate" and refers to how often pitchers strand runners on base to end an inning (they have to be runners that they allowed on base, not inherited runners).
5. HR/FB rate (Homerun to Fly Ball ratio): This simply tells us how often a pitcher gives a home run on a fly ball (not all home runs are captured by this statistic as some are line drives).

2021 Negative Regression Candidates

Using 2021 data only, I did some stat dives this weekend to try to identify some of the pitchers whose ERAs aren't supported by the rest of the numbers. I pulled the data on all starters who have pitched at least 30 innings so far this season. Here's our first chart, the top-12 pitchers sorted by the difference between their ERA and SIERA.

Note: The data displayed here was generated BEFORE Sunday's games and doesn't reflect Matt Boyd's or Marcus Stroman's most recent starts.

At first glance, you'll a lot of pretty solid pitchers and there are definitely some guys here who are capable of keeping up their strong starts, too. The three main stats that I analyzed were LOB%, BABIP, and HR/FB% and I compared every starters' current rates to the league averages to find the difference.

In my mind, these are stats that are all bound to regress towards the average over the course of the season and are largely outside of the pitchers' control. There is perhaps an argument that a pitcher with a high LOB% rate could maintain that because they are really good at buckling down and getting that third out in an inning with men on base. But even the most clutch pitchers can't strike everyone out with two outs and men on base and some of those runners are going to score (sorry John Means, but 100% just isn't going to hold up all year).

BABIP is possibly the best stat to look at as far as which pitchers have been lucky or unlucky as far as getting outs when opposing hitters put the ball in play. The league average right now is .278 and by the end of the season, the vast majority of pitchers should have a BABIP within a few points of that number in either direction.

When looking at HR/FB rate, the league average right now is 13.5%. A really low number should concern us the most with fly-ball pitchers like Matthew Boyd and Danny Duffy. When a pitcher is giving up 35-40% fly balls and so few of them are going out of the park it's a major red flag. Of the two, I think Duffy has the best chance of keeping his renaissance season going (and yes, I have been touting Duffy weekly this season) but I also recognize that there's no way he's going to keep being as good as he's been so far.

Boyd is a guy I would stack against in DFS (and I did yesterday) as I think his BABIP and HR/FB are both highly unsustainable. He gave up four earned runs on six hits but did manage to strike out eight Cubs in the process. He didn't get blown up, but I would bet on those home runs coming sooner than later and will keep targeting him with hitters.

Let's briefly discuss some of the guys we shouldn't be panicking about from this list. Trevor Rogers is the real deal and there's no reason to sell high on him or target him in DFS. Lance Lynn is doing what he's done now for a few years in a row and nothing stands out here to me as being that big of an outlier for him. Carlos Rodon finally came back to Earth in his last start and while his BABIP and LOB% should regress a bit, he's still looked great with the added velocity and strikeouts. I already mentioned Duffy as a guy who can still be pretty good going forward even if he gives up a few homers.

The one really good pitcher who stands out here is Means. He's never done anything close to what he's done over his first eight starts. The no-hitter was awesome to see and I've always liked Means and his awesome changeup. But are we really looking at a 28-year old breakout year from Means and Cy Young type numbers? The 100% LOB rate and .182 BABIP really stand out as major outliers and will absolutely regress.

A quick look at his Statcast sliders show that's definitely been good almost across the board, but not 1.21 ERA-level elite as we would need to see. As a guy who throws about 92-93 with his fastball, he doesn't have much margin for error with his pitch location and we can see that he has had some balls hit really hard already this year when you look at those max exit velocity and barrel % numbers. The good news is that even with some regression, Means is still projecting as a really good pitcher - just not a great one.

Now let's look at our poster child for negative regression - John Gant! A few of my fellow Rotoballer MLB writers have been complaining lately about Gant and how he continues to somehow avoid the blow up outing. For those who have been frustrated by stacking against him without getting much in return, take heart! His 1.83 ERA is entirely smoke and mirrors! With a SIERA that is four points higher, he blows away everyone else on our leaderboard.

The crazy thing is that other than the LOB%, none of his other numbers really stand out on my chart. But we are talking about a guy with an 18% K rate this year and a whopping 17% walk rate. He is constantly allowing baserunners but somehow avoiding the bad start, allowing three earned runs or fewer in all seven starts this season. Let's take a closer look at his Statcast data.

Yikes! I think it's safe to say that John Gant is not a good pitcher and that he has the best chance of seeing major negative regression of the guys on our list. Keep stacking against him, he simply can't keep this up!

2021 Positive Regression Candidates

Note: This data doesn't include data from Sunday's games and Kyle Hendricks' most recent start.

I want to touch on some of the guys who shouldn't be as bad as they've been, too. I think we can all agree that Luis Castillo is not a 7+ ERA pitcher, that much is obvious. But he's not been good, either and his SIERA suggests that even if he had better luck with stranding runners (that 51% LOB rate is crazy low) or with balls in play (.380 BABIP also really high) he would still be underperforming his preseason expectations.

Dylan Bundy is a guy I wouldn't be bailing on and I expect him to bounce back and have a productive year. His 57% LOB rate has plenty of room for positive regression and nothing else really looks out of place. Andrew Heaney is another guy who has pitched pretty well this year despite a 4.75 ERA. He's a HR-prone pitcher, but a 20% HR/FB ratio suggests he's given up a few too many than he probably should have. Jameson Taillon is another guy who's probably going to be better than his 5+ ERA suggests, but there are no glaring differences here to suggest why he'll improve.

I wouldn't count on Kyle Hendricks morphing back into a top-end starter, but he did throw eight solid innings yesterday (albeit against the Tigers) and has been largely a victim of the home run this season with a massive 28% HR/FB rate. His Statcast numbers are still pretty bleak, but his control and quality offspeed pitches give him at least a fighting chance at being serviceable despite his complete lack of velocity.

The rest of these guys I would lump into the category of "bad pitchers who might not continue to be THAT bad" with Jorge Lopez being the poster child. Lopez has given up a lot of homers this year but has pitched better in his last few starts. I have stacked against him unsuccessfully already a few years and should probably rethink that going forward. Jordan Lyles and Trevor Williams both usually trigger my "must stack against them" reflex, too, but both guys can troll us at any point - they're just not all that terrible.

Matt Shoemaker might have a category of his own - "pitchers who have been horrible and might only be pretty bad." Keep picking on Shoe.

Well, that's all I have for you today! I hope this information helps you make more informed decisions this season and make sure that you continue to follow our MLB DFS strategy articles throughout the season!

More Fantasy Baseball Analysis