May 16, 2003
Balls, Strikes, and the Reds' Win StreakKen Griffey's back in the lineup, Ryan Dempster's heading back to the rotation, and Adam Dunn and Austin Kearns are each on pace to hit 53 homers. With their win over the Cardinals on Wednesday, the Reds moved into a tie for second place in the NL Central. They have recovered from their 0-3 start, when the Pirates outscored them 24-10, and though no one could expect their torrid streak to last, the Reds are being talked up as a legitimate playoff contender.
This is where the Pythagorean record comes in handy. The Reds are actually 21-19, but their expected record is 16-24; because no matter how good their hitting has been, their pitching has been much worse. Going forward, the expected record is the better indication of how the Reds will finish the season.
As is often true in baseball, the ball-strike count plays a big part.
The Reds lead the league in home runs, but they also lead in home runs allowed. Of all teams not playing half their games in Denver, Cincinnati has the league's highest ERA. They have allowed the most runs and most walks, and only two teams--Colorado and Milwaukee--have given up more hits. The Reds' staff has racked up the fewest strikeouts, and their starters have the league's highest rotation ERA by far, more than a half-run higher than San Diego's. It's why Cincinnati's Pythagorean record is only 16-24.
Why have Cincinnati's pitchers done poorly? The ball-strike count plays a big part. In a previous column, I noted that the Reds' strike rate was tracking with their wins and losses. This is an essay about the importance of strike rates, but let me get the caveat out of the way: Strike rates play a big part, but of course they don't explain everything. High strike rates don't necessarily mean success. Brian Anderson rarely gives up walks, but he gets hit hard. That said, strike rates have a big influence on outcomes. Pitchers and batters alter their approach to an at-bat on where they are in the count. Whether or not a pitcher works ahead in the count and can use his whole repertoire matters, and it matters whether a batter has to protect the plate because he's behind in the count. Strikes reflect how well a pitcher controls a game.
What might surprise a lot of people is that there is no official method of scoring balls and strikes, just as there is no official way to publish a box score. You won't find anything in chapter 10 of the official rules on box score formats or ball-strike tabulations. There is no official method, but for both box scores and pitch scoring there is a customary, standard way of doing things.
Under the standard, swinging strikes, called strikes, foul balls on full counts, and balls in play are counted as Strikes. All balls, including those thrown for pitchouts and intentional walks, are counted as Balls. As strikes are recorded now, a called third strike is no different than a 500-foot homer. An intentional ball is no different than a wild pitch.
Various changes to this method have been proposed, and in that previous column I suggested that "bad strikes" and "intentional balls" should be counted differently than they are under the standard. A 500-foot homer should not, in my view, be counted as a strike. The argument against this is that hitters sometimes hit good pitches, most hits are off pitches in the strike zone, and pitchers shouldn't be penalized for it.
The emphasis shouldn't be on merely finding the strike zone, but on using the strike zone well. It's not enough just to "throw strikes." The strategy of getting ahead in the count is nothing new. Some will argue that pitchers are already trying to throw strikes, especially first-pitch strikes, and so what we're seeing is a result of that strategy. That might be true for some pitchers, but one of the most common complaints among analysts, broadcasters, and former players is that today's pitchers are timid. Not everyone is trying to throw strike one. There is a large number of marginal pitchers in the majors. As Gary Gillette points out, some of these pitchers are nibbling because they have minus stuff and know they'll get hit hard if they put it over the plate, and there are other pitchers who are unthinkingly taking the advice of everyone who says that all you have to do is throw strike one.
If you become predictable on the first pitch, batters will adjust and start hacking. But if you have movement and location and change speeds, if you aren't one of the marginals attacking in the zone shouldn't be a problem even on the first pitch. Tim McCarver has it right when he says that when broadcasters tell viewers that pitchers need to throw strikes to stay ahead in the count, it would be more accurate to say that pitchers need to stay ahead in the count by getting strikes. As he points out, why throw the ball in the zone if you can get the hitter to chase a ball in the dirt? He estimates that 80 percent of swinging third strikes are on pitches out of the zone. If you have empirical evidence to disprove that, let me know. His estimate might be inflated but his point is well-taken. So there should be no problem rewarding a pitcher for getting a batter to chase pitches outside the strike zone, but rewarding him with a strike for giving up a hit is too generous for my taste. I don't feel sorry for them. Don't they get credit when a batter chases a bad pitch and makes an out?
You've gotta throw the ball over the goddamn plate and give the other man his chance.
Leonard Koppett suggests that balls hit be scored separately from strikes, but I think that goes too far. It's fair to credit a pitcher with a strike when he allows a ball in play that ends up as an out. That's the desired effect, isn't it? You don't want strikeouts, and you don't want the batter hanging in there working the count. You want outs early in the at-bat. Tom House notes in his video The Pitching Edge, that since batters make outs more than 70% of the time, the odds are with you if you throw the ball over the plate and make the hitter swing the bat. "Throw the ball over the plate and let your teammates help you on defense," says House. "You don't try to make the hitter miss until he has two strikes." House overstates the case a bit, since he ignores walks. Batters fail 6.5 out of 10 times, not 7. And while House's point about throwing strikes is valid early in the count, Gillette says:
"Pitchers are afraid to throw strikes in hitter's counts. If pitchers consistently came into the hitter when behind in the count, there would be a lot fewer walks and much higher batting averages. This effect can be seen when looking at walk rates and batting averages with the bases empty and with runners on base. With the bases empty, pitchers are more willing to try for the quality strike (at the knees or on the corners), and they accept more walks as part of that trade. With a runner on first base, pitchers are loathe to walk the batter and will accept a greater chance of a base hit."
But House is right that it would be ideal to induce poorly hit balls rather than try to rack up strikeouts. In principle, counting outs-on-balls-in-play as strikes seems fair to me.
Mentally the computation is easy. To see how this works, let's look at a hypothetical game, where the total pitch count is 120 and the team gave up a lot of hits:
Pitches/Strikes 120/73 Hits - 11 ---- Good Strikes 62
And it would be unfair to count intentional balls the same as wild ones, so we need to deduct those from the pitch count total. STATS and Baseball Info Solutions do this already when they score games. Box scores report intentional walks but not pitchouts. Since we're trying to improve the method here, not perfect it, intentional walks get us close enough.
Pitches 120 Intentional Balls - 4 ----- Revised Pitch Total 116
Under the standard method, the strike rate would be 120/73, or 61%. The revised strike rate would be 116/62, or 53%. What does the difference show?
An average strike rate according to standard practices is 61%. If you take out the hits and account for intentional balls, an average strike rate is 57%. In the example, the pitcher had an average day in the strike zone. Under the revised method, he was below average because he gave up a lot of hits. It's not a perfect way to look at it, but it's easy to do and it just seems fairer to me.
The revised method makes the Reds' pitching woes more obvious than the standard method can. Under the standard method the Reds have had 18 games where they have attained at least an average strike rate. Account for the hits and they have made the mark only nine times. They are near the bottom of the league in hits allowed, and this explains the difference.
Their pitch counts don't represent a lot of third strikes. The Reds throw a lot of balls. They walk a lot of batters. And they work from behind in the count too often and find themselves having to groove pitches just to avoid walks. While it might be true that a pitcher has little influence over whether a ball in play--other than a home run--becomes a hit or an out, he does have some say over what kind of balls will be put in play. As the count favors the batter, his likelihood of hitting a double or triple goes up. In cripple counts he's looking for pitches to drive, so this is what we'd expect. There's a huge difference between an extra-base hit and a single, so it does matter whether a pitcher gets ahead in the count and gets outs early in the at bat. Throwing strikes--getting strikes--is important, and the strike rate is relevant to that.
Either the Reds are wild or they nibble too much and get themselves into trouble. During the nine-game hot streak they have had a substandard strike rate six times, all but the last three games. It's not enough to say that when their strike rate is poor the Reds lose and when it's good they win. The strike rate is just an indicator of how their pitchers are controlling the games. The Reds might survive a poor strike rate in a single game, or even a series of games, but over time it will place an unsustainable burden on their bullpen--the Reds starters have been treated gently but their relievers have been worked hard--and place impossible scoring demands on their lineup. Nate Silver projected the Reds to finish two games under .500. Right now they're two games over. If they finish over .500 without improving their strike rate substantially, I'll set my hair on fire.
The italicized section headings in this column are taken from Weaver on Strategy, written by Earl Weaver and Terry Pluto, and edited by Chris Kahrl. I also want to thank Gary Gillette and Keith Woolner, who were kind enough to do a sanity check on my arguments. The opinions here are no fault of theirs.