The Year the Orioles Wouldn’t Flinch

There are certain statistics that get tallied on scorecards such as hits, walks, runs, and strikeouts.  These aren’t the modern, comprehensive measures of performance we typically use here at FanGraphs, but they are the building blocks of those metrics. Without singles/doubles/triples/etc, there is no way to build wOBA or wRC+. Without strikeouts, walks, home runs, and innings pitched, there is no FIP.

While we’ve generally moved beyond caring about certain counting stats, we use them to build the things about which we care a great deal. But there is at least one standard scorecard event that doesn’t get a lot of attention when we build these metrics because it’s extremely rare and also extremely subjective: the balk.

The layman’s description of a balk is easy enough to understand: it’s a movement made by the pitcher intended to deceive the runner into thinking that same pitcher is about to throw towards home, when in reality he (the pitcher) is not. Any reasonably informed fan knows that the specifics of the balk rule are complicated and enforced at the whims of the umpire. It’s a judgment call, and one that doesn’t seem to be uniformly implemented at any given moment in time. But there are trends in balking:

This graph deserves two explanations. First, let’s discuss 1988. The balk rule was changed in 1988 and if you didn’t know the results of the change, the actual difference in the wording of the rule might not catch your eye. I’ll allow Theron Schultz of Recondite Baseball to explain:

Baseball Official Rule 8.01(b): The pitcher, following his stretch, must (a) hold the ball in both hands in front of his body and (b) come to a complete stop.

1988 Baseball Official Rule 8.01(b): The pitcher, following his stretch, must (a) hold the ball in both hands in front of his body, and (b) come to a single complete and discernible stop, with both feet on the ground.

The difference between the two rules is that the 1988 version replaced “complete stop” with “single complete and discernible stop, with both feet on the ground.” This slight change, intended to make balk calls more uniform throughout major league baseball, instead sparked one of most frustrating summers ever for major league hurlers. Only six weeks after opening day, Rick Mahler of the Atlanta Braves committed the 357th balk of the 1988 season, breaking the MLB record for most balks in a complete season… with three-quarters of the season to play. Before all was said and done, American League pitchers were called for a staggering 558 balks. Their National League brethren had it a little easier, “only” committing 366 balks.

This is a wonderful bit of trivia, but it obscures the broader trend: fewer balks are being called across the league. We have data back to 1974, and since that time, there has been a clear decline in the number of balks called in major-league baseball (shown in the graph as balks per 162 games). Here’s the same graph a before, now with a truncated y-axis for easier viewing:

It’s unclear if umpires are letting pitchers get away with more balking behavior or if pitchers are less guilty than they used to be, but the trend is rather clear. There was also a rule revision before 2013 outlawing the fake-to-third-throw-to-first move, which shaved about a balk per team season from the league.

One complication for any sort of balknalysis is that we’re dealing with an incredibly rare event. In 2015, the average club balked fewer than five times, making it pretty hard to develop any sort of predictive model of balking or retroactive model of value. Obviously, balking is a negative even for the defense and a positive one for the offense, but the base-out state can swing the value pretty heavily one way or another and we don’t really have enough data to get a good average in any one season. You can calculate it, of course, but there’s a lot of noise and not much signal.

If individual pitchers balked 10-15 times a year, it might be worth building it into our WAR models, but when a mere two or three balks represents a bad season for a pitcher, it’s hard to find reason to over-complicate things.

But that doesn’t have to stop us from noticing oddities. Let’s start by getting an idea of the distribution of balks per team per season since 1974. I’ve lopped off the 1988 season because the rule was different and it makes the graph harder to read:

One thing you likely notice is that while balking is rare, it’s also extremely uncommon for a team to complete an entire season without balking. It’s happened just nine times in recorded history. The teams to do it were the 1975 Red Sox, 1998 Tigers, 1999 Red Sox, 1999 Astros, 2003 Red Sox, 2003 Yankees, 2007 Red Sox, 2007 White Sox, and… the 2015 Orioles!

My brain, and probably yours, wants to make something of the Red Sox appearing on this list four times, but it’s pretty hard to come up with a reason for why an organization would have some sort of anti-balking secret that wouldn’t eventually spill out into the rest of the game. It’s also interesting that these teams seem to cluster, with three years featuring two clubs avoiding balks altogether.

It’s not surprising at all that there would be more no-balk teams as time moves forward because there are fewer balks in general, making it more likely that any one team will go an entire season without balking simply due to random variation.

For reference, the standard deviation for balks per team season is 4.8 since 1974 (excluding 1988) and 2.9 since 1995. Roughly speaking, if balking were totally random and normally distributed since 1995 (in reality, it’s slightly skewed), you would expect to see a zero in about one out of every 20 team seasons. There have been zero-balk seasons in eight of 624 team seasons (1.3% of the time).

As I noted earlier, there isn’t going to be a way to definitively prove that a team has low-balk skill because the frequency of balks is so low and the range is so compressed, but I am curious to see if we can point to any anecdotal evidence that might suggest that goose egg produced by the Orioles was something more than totally random.

My first thought was to check pitcher handedness, as one side might open themselves up to balks more often than others. The 2015 Orioles threw from the right side more often than the average team, coming in at 76% of batters faced from the right side compared to the league average of 73%. Lefties certainly get a lot more gray area in their movements, but they are also probably more open to balk calls due to the direction they face while on the rubber.

Orioles pitchers were 15th in the league in terms of total plate appearances with men on base, though, so it’s not like they were simply in fewer balking situations than average. Additionally, I grabbed the average number of balks per 1,458 innings pitched (162 games season, give or take) for the 2015 Orioles pitchers across all of their careers entering 2015. The collective average was 4.2 balks, making them statistically no different from average.

Finally, balking seems like something a pitching coach might be able to train out of you. Looking back at Dave Wallace’s career as a pitching coach, we find that his teams actually do balk slightly less than average. From 1999 to 2014, teams balked a little more than five times per year, while Wallace-led staffs balked just under three times per year. This might seem meaningful, but it’s worth pointing out he took over the zero-balk Red Sox in 2003 midway through the year and that it’s virtually impossible to track when and where a pitcher may or may not have learned to limit his balking.

For example, was Wallace responsible for the no-balk 2007 Red Sox even though he had moved on to Houston after the 2006 season? With something as narrow and specific as this, it would be awfully hard to shower him with praise for something that could simply be noise. He also spent time as a minor-league pitching coordinator over the span, which reminds us how many people influence any one player.

Ultimately, the Orioles were a slightly right-handed club with a pitching coach who may have a slight downward effect on balking, but there’s a good chance that the 2015 Orioles simply happened to not balk for a stretch of 162 games. Although, it is worth noting the oddity that zero-balk teams are more rare than we would generally expect given the mean and standard deviation of recent history.

Balks are on the decline across the league, but that somehow hasn’t led to a similar increase in zero-balk seasons. While the Orioles’ balkless season was likely something of a fluke, it was probably a delight for die-hard scorekeepers who had been tracking the feat all along.