Seattle and Baserunning

The Seattle Mariners’ inability to score runs is no secret. Entering Sunday’s affair with the Los Angeles Angels of Anaheim, the club had played 91 games and scored 309 runs; a 3.4 runs per game average. The Mariners scored two runs (and won) but it took extra innings. So much attention has been paid (and rightfully so) to how Seattle prevents runs and not enough on how they score them – or rather, how they could increase their scoring without resorting to sinning such as rosterbation or promoting Dustin Ackley before he’s ready.

One of those ways is taking more risks on the basepaths. Not just stealing bases but the other aspects of risk in baserunning; namely attempting to take extra bases on hits. Take one of the most common situations in the game: a single with a runner on second base. As the runner runs the 90 foot path between second and third, the third base coach must make a quick calculation on ball placement, runner speed, the fielder’s arm strength and accuracy, the score, and the base-out state before deciding to put on the brakes or send the runner barreling home.

One of the ways we analyze when (or if) the Mariners should be taking risks on the basepaths is by looking at run expectancy charts and empirical data. Using a Markov Chain to generate a run expectancy table we can then figure out the break-even points for various out states surrounding the aforementioned situation. The Markov Chain gives us a BaseRuns estimated 3.9 runs per game figure for Seattle, one that actually happens to be on the high side when faced with reality. That’s because the Markov Chain doesn’t account for outs on the bases – it’s just not designed to do so. This is not a projection and there are no regression or adjustments made to the team’s numbers. Here is the generated RE chart for the Mariners to date:

They’re a low scoring team in a low scoring environment. Simply put: that means that making an out isn’t as taboo as it would be in a higher run scoring environment (almost anywhere else in baseball) because it would be unlikely that the runner scored anyways. On one hand, this means the break-even point for going second to home is lower than it would be for almost any other team in baseball. On the other, well, they aren’t scoring many runs. That’s a problem.

What I decided to focus on – since I only wanted a snapshot of the Mariners’ baserunning tactic rather than the entire montage – is when a single was hit and a runner was on second base. That doesn’t mean that baserunner was the only person on base. It also doesn’t mean all singles are created equal. Obviously there isn’t always a normal distribution of batted ball types or throwing arms in situations like these, but pretend for a minute there is. What we have below is each of the 12 situations where a runner could be on second base with a single hit as dictated by the state of outs recorded and other runners on. Those states are backed by the Mariners’ seasonal numbers through Saturday’s game; a quick key: Sent means the runner on second was sent home; Stayed means that runner was held on third; and Out means out at home. Simple enough, right?

The league average for sending the runner home is 63% — the Mariners wave runners home 61% of the time – while the league average success rate is around 93% — the Mariners’ success rate is a league worst 84%. As you can see from the above data they’ve made four of their eight outs with two away, good for a 50% rate. They’ve only made one out with nobody away (this play actually resulted in Ichiro being thrown out at home, which seems fairly uncommon) which matches with inherent rule of thumb.

Let’s look at this by success rate and break-even point, though, to see whether the Mariners should actually be running more on account of great success. Now, again, this is based on small sample size and is entirely discounting the throwing arm and batted ball type. It’s not a 100% perfect analysis by any means, it’s just a general overview of the situations at hand:

Based on this analysis, it seems the Mariners aren’t doing so bad when it comes to running after all. With nobody out third base coach Lee Tinsley is waving around 48% of the runners from second; with one out that number raises to 52%; and with two out that jolts up to 73%. Whether the Mariners discuss this with Tinsley or not — he’s actually their second third base coach on the season –, he seemingly has an understanding of the run expectancy matrix and how break-even points lower as the out count rises. As far as success rates per out; 90% of the runners sent with nobody out are safe; 82% with one out are safe; and 86% with two outs are safe. Again, Tinsley seems to get it. If anything, he might too passive with one and two outs, as all the break-even point supplied suggest he could get away with sending more runners, even if a few make outs.

So what of the league worst success rate? Well, remember, that isn’t broken down by outs or base situation. A team could perfectly understand that a success rate of around 50% with two outs is worth their gamble and have that take up 60-65% of their total attempts, thus skewing their overall results to look worse than it contextually is. The same is true for a team that just refuses to allow their guys to go unless there’s a 95% chance of success. Yes, it may look better on the surface, but in reality, that team is probably costing themselves a chance at some more runs by being too loss-adverse.

Note: Inspiration for this piece came from a Russell Carleton ESPN.com article and Jeff Zimmerman’s bit on similar issues with the Royals.