Yankees spring training: Trying to find meaning in small sample size statistics

Most baseball fans know that spring training stats mean next to nothing. The sample sizes are tiny and opposing rosters are often filled with guys from Double-A. This explains how no-name players like Jon Weber and Jorge Vazquez can put up OPS's north of 1.000 in March, only to go back to being minor league fodder once the calendar turns to April.

Nonetheless, there are certain statistics that are more meaningful than others in small samples. For both batters and pitchers, it's been established that stats like strikeout rate and walk rate -- which don't depend on lucky bounces or the quality of defense -- take the least time to become reliable. If any spring training stats were to have any predictive value, it would probably be these. This isn't to say other stats can't be predictive of a player's performance, but they tend to be all over the place across such a small number of games.

Of course, simply looking at these stats doesn't take into account the quality of competition they faced. No matter how well a player performs, it's pretty much meaningless if it comes entirely against players from A-ball. Luckily, Baseball-Reference puts out a stat that attempts to quantify the quality of a player's performance. The stat assigns each player a number from 1-10 representing his average competition: 7 = Double-A, 8= Triple-A, and a score of 10 means the player faced only big-leaguers.

Russell Carleton of Baseball Prospectus estimates that walk and strikeout rates start to become somewhat reliable at the following thresholds:

Hitter K%: 60 Plate Appearances

Hitter B% 120 Plate Appearances

Pitcher K%: 70 Batters Faced

Pitcher K%: 170 Batters Faced

Most players fall short of these totals in just one month of spring training, so I've included all Yankees who made it half-way to the strikeout thresholds, along with the average level of competition they faced.

So based on all of this data, these are the Yankees I reckon have the best chance of over-performing their projections based purely on their spring training stats:

Ichiro Suzuki:

K%: 4% BB%: 6% Opponent quality: 9.3 (Quad-A)

On the surface, it looks like Suzuki had a terrible spring. His .240/.283/.280 batting line is significantly worse than you'd expect from Ichiro, even following his dismal 2013 campaign. Still, Ichiro faced mostly major league pitching and the underlying numbers suggest he was actually pretty decent. Suzuki walked more than he struck out this spring; and although the results weren't there, his .250 BABIP suggests he got a little unlucky. This isn't to say Ichiro was good this spring, but his peripherals suggest he was significantly better than his .283/.317/.372 Steamer projection.

Ivan Nova:

K%: 26% BB%: 3% Opponent quality: 9.2 (Quad-A)

Nova posted a respectable 3.66 ERA in camp this year, but his peripherals show that he was much better than that. Nova's FIP was a sparkling 1.25 this spring, propelled by a 21:2 strikeout to walk ratio. That's about as good as it gets. His performance is even more impressive considering the majority of his completion was big-league caliber.

And the under-performers:

Brian McCann:

K%: 22% BB%: 8% Opponent quality: 9.1 (Quad-A)

McCann's .200/.265/.333 triple slash pretty much tells the story. The Bombers' newly minted catcher struck out in nearly a quarter of his trips to the plate and walked less than he usually does. The only thing falling in McCann's favor is that he faced his share of major league pitching: The average quality of his competition was slightly closer to MLB than Triple-A.

Danny Burawa:

K%: 5% BB%: 8% Opponent quality: 8.8 (Quad-A)

Although he had an impressive 1.93 ERA this spring, Burawa didn't really pitch well at all. The hard-throwing reliever walked more batters than he struck out, earning him a 4.96 FIP. His opponent quality also wasn't great and suggests his average opponent mirrored an average Triple-A player. Burawa got really lucky in his 9.1 innings this March, which allowed him to post a sub-two ERA. Otherwise, there's not a lot to like.

And loud springs that may not be as great as they look:

Yangervis Solarte:

K%: 15% BB%: 11% Opponent quality: 8.1 (Triple-A)

Solarte went H.A.M. this spring. The little-known Non-Roster Invitee hit a disgusting .429/.489/.571, which was enough to land him a spot on the Yankees' Opening Day roster. Solarte certainly hit the snot out of the ball, but there are reasons to be skeptical going forward -- even beyond the typical spring training caveats. Solarte actually put the ball in play less than he usually does: He struck out three percentage points higher than he did in the minors last year. More than anything, his spring performance was driven by lucky bounces, as evidenced by his .457 BABIP. And lastly, Solarte had the lowest opponent quality score of any Yankee in camp. He basically did what he did against Triple-A pitching.

Dellin Betances:

K%: 23% BB%: 8% Opponent quality: 8.6 (Triple-A)

Like Solarte, Betances impressed enough this spring to earn a niche on the Opening Day squad. He posted a 0.73 ERA in 12.1 innings, but his 11:4 strikeout to walk ratio wasn't quite as impressive. On top of that, his opponent quality was one of the lowest on the team. There's no denying that Betances had a very good spring, but based on his strikeouts and walks, he wasn't as uber-dominant as his ERA implies.

Admittedly, this was probably a frivolous exercise. Pretty much anything that happens on a baseball diamond in the month of March holds very little water and your opinion on these players should be more or less the same as it was two months ago. Nonetheless, if I had to identify players who "turned the corner" based purely on spring stats, Ivan Nova and Ichiro Suzuki would be my bets.