Coming off an eight-game winning streak, it’s easy to forget where we are in the season (four games below .500 and 44 runs in the hole, by the way). During a single season, every stat—from batting average to OBP to runs scored to ERA to xFIP—jumps around a lot, even after 102 games. Below I have compiled time series graphs to monitor the ups, the downs , and the stabilization of Seattle’s primary stats throughout the season.
Above you see the Mariners average runs scored (green) and average runs allowed (red) through each game during the season. As the season has worn on, these averages have begun to stabilize because the average of five games is easily swayed while the average of 100 games is hard to budge, even with a 10 or 15-run performance. Since about game 75, the M’s offense has been hot, but not hot enough to catch up to their runs allowed averages which have also been increasing. It kinda makes that eight-game winning streak (where we outscored opponents 50-to-24) seem small and insignificant. I’m sorry about that.
In the first graph, we saw that the M’s were allowing an increasing number of runs in the past month. Though it’s tempting to blame that on the starting rotation, the general xFIP trend has been pretty steady, as shown by the blue line. In this graph, I took each one of Felix Hernandez’s starts, and then took the year-to-date average xFIP* of him and the four starters that followed him. The spikes are due to weeks when a new pitcher entered the rotation (like Hector Noesi or Erasmo Ramirez), and the big dip after Felix’s 20th start is because Iwakuma got to start twice in only a few games due to the All-star break. Overall, there has been no obvious trend in any direction since early on this season.
The Mariners’ bullpen got off to a rocky start in the xFIP department, then pitched well to about the 50-game mark, and has since leveled out to an ERA equivalent of about 3.70. xFIP is perhaps not as clairvoyant for bullpens as starting rotations, but it suggests that the Mariners’ 4.5 bullpen ERA might come down.
Because apparently UZR is not updated on a daily basis, the UZR time series above often gets stuck on the same number for up to a week at a time. That doesn’t change the fact that the Mariners have been playing defense progressively worse by UZR standards. Remember that each data point on the time series is cumulative, and thus the fact that the M’s have managed to go from a –7.5 around game 52 to a –11.6 currently is impressive. Impressively bad.
It would be a misinterpretation of the graphs to assume that the positive offensive trend and the negative defensive trend will both continue on their respective paths. Because the graphs are cumulative, it more likely that they will start to plateau. In fact, Fangraphs projects the M’s to score about 4.0 runs the rest of the season—right where they are now.
Every season treats us to ups and downs, but the long term averages tend to lie less often the true makeup of a team. The Mariners are still a below-average team, but less of one than a month ago.
*xFIP here is a modified version of Fangraphs’ linear weights formula. There are slight discrepancies between Fangraphs’ version and mine, but they are both regressing ERA to the mean based on walk and strikeout rates. My formula then utilizes groundball rates as the third componenty because, well, that’s what I had available, and Fangraphs uses flyball rates.