Here’s an amusing little chart I just made:
It’s a seven game moving average of the Rangers runs scored (blue) and conceded (red) this season. Click on it to enlarge. What’s the obvious thing you see? Of course, it’s the giant spike at the right hand end, showing what a freak show the 30 run game was, how much it affected things. It was a crazy score, something you will almost certainly never see again in your lifetime, something even my two year old will probably never see again. It was a once in a century score. And with moving average, we see just how far out of whack with the rest of the Rangers games this season it was. If I waited until the end of the season to show this, you would see that line drop way back down again, an extreme spike in an otherwise moribund year.
But we can use this chart educationally, to look at just how the Rangers have been doing. Using a moving average eliminates much of the jaggedness of the chart, giving slightly lower peaks and valleys, allowing you to see trends more easily. Now, comparing runs for and against, you can kind of follow along with the season. Long periods where the blue is below the red are the losing times, and where blue is above are winning times. If they run fairly even, the team should be about .500. You can see for most of the first half of the season, the Rangers were below, for a short time they were above, and then back to even and below again. You’d expect to see that reflected in winning, and in general you do.
There are however a few interesting things that the moving average does not eliminate, which includes excessive wins. Obviously the 30 run win obscures a lot of bad stuff around it, you can see the Rangers were falling into a hole approaching the worst of the season when they exploded (and everyone knows they’d been bad in the previous two games). Ignoring that though, May immediately sticks out to me. Remember May? It was the worst of times. The Rangers sucked hard in May, going 9-20 on the month, but looking at the chart they almost seemed decent. What’s going on? In May they scored 145 and conceded 160, which Pythagorean wins would show as a 13-16 month, not 9-20. In fact in May they had three wins of more than 10 runs, and apart from a 9 run loss their next biggest loss was by 6 runs. Those three big wins accounted for 42 of their runs for and 7 runs against, which is a huge margin in just three games. Take them out and you see they’re down to 103-157, which moves them back to a Pythag of 9-17, much closer to what they truly were. For May, they didn’t lose big, they just lost a lot of little games and the three big wins masked the trouble more than they should have.
Looking a little earlier, April was another bad month, barely above water all the way. You can in fact show that through mid-June, the Rangers averaged about 5 runs for and 6 runs against, which is a deadly combination when you’re trying to kick off your season (the actual averages through June 15 were 4.94 for and 5.86 against). At that point it’s even hard to decide who to blame. The offense seems decent at 5 runs per game, pitching is of course bad near 6, but even then I remember complaining a lot about how the offense was struggling. I suspect it is consistency that is the key, meaning that they never got in a stretch where you felt they would get five runs every night, but more a week of two or three runs then one game of 15 to balance it out, as they did in May. At least the pitching was consistently bad.
The high point of the year was late June to late July, a period when they went 23-14 and things seemed to be working. The chart shows the second half of June as being the best of times, when the offense was much the best of the year, and the pitching was running well too. That two to three run gap turned into a lot of wins, but was unfortunately a very short streak. While the pitching kept going well through July, the offense fell away to the same level, leading to a 14-12 July in which the team was outscored by 15 runs. Slow increases on both sides ended up with the pitching going higher and hitting going lower in August, leaving the team (and it’s fans) flat once again. Interestingly, as a whole, both the Rangers and their opponents combined have scored a lot fewer runs per game in the second half. 10.8 runs per game through June 30, 8.3 from July 1 to August 21 (just over 9 through today, which again shows the large impact of 30 runs in one game, it raised the average for two months by 0.7 runs per game). Everyone always says the Rangers start well but fall apart when it gets hot, is this a sign of that? Not just on the Rangers, but their opponents too, scoring 2.5 runs per game fewer between them. Good pitching and bad hitting, something you would never expect to hear about a Rangers team.
This almost turned into a season review kind of thing, which was not what I intended. I was simply trying to show how far out the 30 run game could throw things. I ought to just copy and paste most of this once we get round to October and I do a real review. Now I’ll have to think of something else to say.
Hey, something I haven’t researched, but I’d be willing to bet that the 29 hits the Rangers got on Wednesday is a record for a team that was no-hit in the same season (certainly in the modern era, can’t guarantee those pre-1900 teams). If I had a good database, I could find this out, but a cursory look shows that the D-Backs were no-hit last year and also had a 20 hit game. 29 is of course much more rare than 20, but even that goes to show that a no-hitter is another kind of fluke. Now the Rangers have been close to being no-hit a couple of other times this year, which would make the 29 hit game even more improbable, if that’s possible.
Some comments on other things going on lately, specifically Ron Washington. He (and presumably Mark Connor) have badly mismanaged the team this week. I was shocked that Millwood went out for the 9th inning tonight, he had clearly hit a limit, and in fact ended up giving up an insurance run that took the wind out of the sails. He lost a quality start, his Game Score fell from 51 to 43, all of the positives we could have taken out of it were blown away simply because they wanted to get him a complete game. What’s the benefit of a complete game loss? Especially when you have just one strikeout, and 13 hits, to show for it. Millwood was not on today, but struggled throughout, and to keep him out in the 9th was kind of like putting a nail in the coffin. The unfortunate thing for Millwood is that he’s had no run support lately, he’s been pitching well but has nothing to show for it.
The other one was yesterday, Kam Loe’s start. Remembering that Loe has just spent time on the DL, and had just one game back where he only went five innings, how long do you think he should be left in? You’d start at five innings, but you might want to go to six if his pitch count isn’t too high. You’d certainly be aware of keeping a close eye on him when you hit five innings, wouldn’t you? In fact, in the middle of the 6th, when he gives up a leadoff homer, then loads the bases with one out, you’d have someone ready in the bullpen to come in and save the day, especially with a 3-2 lead. On the other hand, you’ve got the 8 and 9 hitters up, and even though he’s thrown 89 pitches already, you want to take another positive out of it. When Jose Lopez singles to make it 3-3, you think about pulling the pitcher again. Even when Betancourt pops out to the catcher, you think about it, because now you’ve got a tired pitcher, bases loaded, and Ichiro coming up. What do you think will happen? I’ll tell you what will happen, in fact I said it right before the at-bat began. Ichiro isn’t going to hit a home run, because he doesn’t have good power, but I predicted he would double and clear the bases, and sure enough he doubled and cleared the bases. Then what happens? You bring in Mike Wood, and he serves up a home run to Vidro on the first pitch. Boom, game gone, anything positive for the pitcher gone.
This would probably be my primary criticism of Ron Washington this year. He leaves pitchers in too long. Is it Mark Connor’s fault? Yes, at least some of it, because it’s his job to tell Ron when it’s time to get the guy out of there. Of course, Connor looks clueless every time you see him, so if you’re trusting in him you’re already in trouble. But ultimately it’s down to Washington, and with all his experience he hasn’t learned when to get a pitcher out of the game. In mitigation, he spent his coaching years in Oakland where they rode their horses into the ground (Hudson, Mulder, Zito), but if he can’t recognize that our pitchers are not of that quality then he has even more issues than knowing when to pull them.
One of the things I don’t have, and am not sure where to find (without creating it myself) is a database of how pitchers performed against their last batter faced in a game, or in the last inning. I’m sure it’s horrendous, since that’s why they’re pulled, but it would be interesting to compare pitchers and/or teams and see how the results stack up. I’d be willing to bet there’s a lot of cases where Washington has allowed people to get into bad situations before getting them out of the game. One method would be to look at ERA by inning for starters, but that doesn’t really tell how far down they got right at the end of their outing. Another would be to see where they were when they left the game, but again doesn’t tell the whole story. If you knew just the numbers on Kam Loe when he left the game yesterday, you’d see that he left with two out in the 6th, runner on 2nd and down by 3. That doesn’t tell you if he is down 0-3, 7-4, or 30-27. It doesn’t tell you if he gave up three runs in the first, then pitched a no-hitter until giving up a double in the 6th and being pulled. Or alternatively if he no-hit them for 5.2 innings, then fell apart. More detailed analysis might help. In particular I’d like to look at the comparison between the start of the last inning pitched, and when he was pulled. In this case, he’d be at 1 run, 6 hits, 4 k’s, in 5 innings, a pretty decent outing. Comparing perhaps the win probability of the start of the inning to when the pitcher was pulled, and see just how much that changed. Then compare that across pitchers, teams, managers, and see just how Washington compares to other teams. That’s a long project, I think, but it’ll go in my list of things to do.
Willie Eyre is going to have Tommy John surgery, and will miss all of 2008, at least. From the start of the season through June 24, he had a 3.00 ERA in 36 innings. On June 26 he had a spot start, and from that point on his ERA was 7.59 in 32 innings. Pretty clear what the two halves of his year were, huh? I won’t say the spot start did him in, he pitched 4.2 innings in that game (69 pitches), on five occasions before that he’d gone at least 3 innings, once even throwing 71 pitches. I remember that for a month before that game I had said he ought to be considered for a start (we were crying out for pitching at that time). I don’t know if it was the workload in that one game, or in the games after, but he had been struggling. He had a second start on August 4, giving up 7 runs in 2.1 innings, which would appear to be a much more likely indicator of trouble. Sorry to see him go, but ultimately the long man spot in the bullpen is one that you can fill easily and cheaply.
The team is in a funk again and so am I. Every game report you read for the next week or two (possibly even the rest of the season) will mention the 30 run game, usually along the lines of “the Rangers haven’t done anything since they scored a record 30 runs”. The chart I started with shows they’ve not been doing anything for almost two months now. I don’t have a clue what to do to get them going. I don’t think they have a clue either.