No Crying in Baseball
August 9, 2006

It all went perfectly to plan... well, almost.

Perhaps the most enjoyable league I participate in each year is the Strat-o-matic experts league known as SOMBOE.  The reason I like it so much is that things like defense and bullpen strength are vitally important aspects of winning, just like the real game.  If it was a keeper league, I might be tempted to not play in any other.  It's that good.

Anyway, I went into this year with an agenda: I wanted to see if the value of stolen bases was more important than recent statistical analysis suggests.  I would build a team with no other outstanding quality other than speed to see if it could hang with the on-base/power teams on offense.

This is the theory: the reason most pitchers parks are regarded as pitcher friendly is because they greatly diminsh the threat of a home run.  Home runs are different from other types of hits because they automatically score a run.  That may sound obvious but it's an important distinction because it's possible to pitch or defend around every other type of offensive event.  With any other type of hit, there's always a chance the run won't score.  So in a pitcher's park where the threat of home run is reduced, it becomes imperative to maximize one's scoring opportunities.  Getting a team of high on base players increases the number of times getting on base, but making a team of basestealers increases the number of times baserunners get into scoring position.  Mine was a choice of quality over quantity, although the two are not mutually exclusive. 

With a few exceptions, most catchers throw out between 25% and 35% of baserunners and the average team gets between 11 and 12 baserunners per game.  If a team is aggressive and tries to steal every time (or nearly every time) they get on base, even if they are only successful 66% of the time, they still advance 7 or 8 extra bases during the course of a game.  They risk 4 or 5 extra outs, but they also increase their chances for scoring 7 or 8 times per game.  With careful tinkering as to when and how often to steal, those percentages can get even better.

So how much do the extra bases matter?  Here are two charts that break down both the likelihood and quantity of runs scored.

Average number of runs scored by number of baserunners and outs:
            0 out  1 out  2 outs
empty       0.56   0.30   0.12
1st         0.95   0.58   0.25
2nd         1.20   0.72   0.34
3rd         1.46   0.99   0.39
1st,2nd     1.57   0.99   0.46
1st,3rd     1.91   1.24   0.54
2nd,3rd     2.09   1.48   0.63
loaded      2.41   1.66   0.81

Percentage Chance a single run will score by number of baserunners and outs:
            0 out  1 out  2 outs
empty       .30     .18    .08
1st         .45     .29    .15
2nd         .64     .42    .22
1st,2nd     .66     .44    .24
3rd         .86     .67    .28
1st,3rd     .88     .66    .29
2nd,3rd     .85     .71    .29
loaded      .89     .69    .34

So if a runner reaches base with no outs and steals second, he increases the chance his team will score a run by 70% over what it was had he stayed at first base and 19% overall.  Even if a team is successful only two out of every three attempts, they are still better off than had they just sat at first waiting for a base hit.  Perhaps not in the overall average number of runs scored but in the chances that a run will score.  Even in the overall number of runs scored, the difference isn't great.  And that's at 66% success rate.  At a 75% success rate, base stealing becomes a serious weapon. 

It turned out to be true in the 100 or so simulation seasons that I ran before the season began.  In the vast majority of them, my Kaufmann speed team ranked in the top third in run scoring despite having only two true power hitters and four regular players with on base percentages under .350.  That might not sound too irregular, but in a 12-team mixed league where the general consensus is that on base percentage is king, it is highly unusual.  True to form, in our actual SOMBOE season the team finished fourth in runs scored.  Eight teams posted a better on base than mine and five out of twelve had better slugging percentages, yet this team scored enough runs to post the best won-loss record (95-67) in the league.  It is my belief that their 290 stolen bases (vs 100 caught stealing) was the difference maker.   During the course of the season I made very few changes (replacing Livan Herandez, Rodrigo Lopez and Lance Cormier with Brian Lawrence, John Halama and Ian Snell) and none to my offense. 

Granted, it's not conclusive proof that speed kills but it does provide some food for thought.

Unfortunately, the story doesn't have a happy ending.  In the playoffs, despite having home field advantage, along with the second best home record and the best road record in the league, my squad lost in the best-out-of-five first round in five games.  The team they lost to, Will Kimmey's Petco team, matched up well with mine with an even distrubution of platoon splits and a solid bullpen.  Still, if his team had a weakness, it was against team speed.  With most of his pitchers either average or worse holding runners and a catching corps that was slightly above average at throwing out runners, I should have, at least in theory, had a decided advantage going in.  Unfortunately, the dice didn't go my way. 

I'm not complaining about losing to a team with an inferior record and one that went 4-8 against mine head-to-head during the regular season.  San Diego did present some match-up problems - Carlos Delgado and Aramis Ramirez were going to get their licks in, he had solid starters and a pretty good back of the bullpen.  And I suspect most "upsets" in the postseason are due far more to match-ups than they are due to something as random as luck.  As an interesting side note, I'm pretty sure my squad had a winning record against every contending team, yet performed woefully against the teams that had the worst records in the league.  Match-up problems - either with the park or the personnel - was the cause.  It was just one of those weird quirks.  Had the WoodChippers (my team) somehow faced either of those teams in the post season I have little doubt that my guys would have been swept away. 

Anyway, my disappointment stems not from losing but from:

a) the ace of my staff, John Smoltz, who led the league in wins going 21-9 during the season, performing exactly opposite of how John Smoltz does in real life in the postseason.  He was awful, giving up 11 earned runs in 14 innings over two playoff starts. 

b) Manny Ramirez, Tony Clark and Brian Roberts totally disappearing in the post season.  Ramirez led the league in homers with 51 and Clark was no slouch with 40 despite playing in a tough park.  Roberts was equally studly in the regular season with 48 doubles, 19 homers and 37 steals.  In the post-season they combined to hit one homer and one double and bat a collective .170.  It's hard to win when your best players disappear at crunch time.  Conversely, San Diego's Delgado and Ramirez combined to hit 9 homers and drive in 20 in their 9 post season contests (San Diego went on to win the championship).  Given how poorly my team performed in its five postseason contests, it's a wonder the series went as long as it did.  But my biggest disappointment was,

c) HAL choosing not to employ what was overwhelmingly and obviously my team's greatest offensive strength - speed - against a team that was one of the league's most susceptible to it in the most critical game of the season.  In the winner-take-all Game 5, my squad did not attempt one single stolen base despite reaching base 11 times.  This was both inexplicable and unconscionable. It'd be like HAL repeatedly bunting with Delgado and Ramirez.

In retrospect, I might have fared better had I started Randy Johnson in Game 1 instead of Smoltz so I could have used him again in Game 5.  His team had been slightly susceptible to lefties, as evidenced by John Halama's remarkable 11-inning, four-hit, 9 K complete game in Game 4 of the playoffs to send it to a Game 5.  But during the regular season, Smoltz had been 3-0 against them with two dominating performances at home, so it really came down to a flip of a coin as to who would be the two-start ace.  And the reality is that no matter how good your team is, you can't win in the playoffs if all of your best players fail to show up on top of HAL refusing to use your team's strengths.

Just for grins, and to discover if there was anything I should or could have done differently, I ran 100 simulations of the postseason.  It's sounds trite, but the results did give me some relief.  With computer generated managers, the championship breakdown went like this:

Will Kimmey's San Diego - 11%
JP Kastner's Washington - 16%
Lawr Michaels' San Francisco - 27%
My KC team - 46%

So maybe there was some bad luck involved.  No crying here, bad luck is part of baseball just as good luck is.  I've maintained for a long time that luck probably plays less a role in the outcomes of games (and seasons) than conventional wisdom maintains, that much of what we attribute to "luck" is merely due to stuff we can see yet can't explain with numbers.  I'm sure that in the days before on base percentage and pitch counts and other modern developments that there were even more surprising results "explained" by "luck". 

I've also maintained that the best team usually wins in the postseason despite the seemingly capricious nature of a short series.  So I'll hold to those tenets and conclude that while I was the victim of a poorly timed slump, a good argument can be made that Will's team was the best.  He started and finished the season on a roll, he matched up well with my squad and the chips fell his way when it counted.  Congrats to him!   I don't know if I'll go with another speed team next year because there really isn't anything left to prove there.  But if you have any unusual ideas or theories that you'd like tested by an expert league, send them to me and I'll put them to action next time around.