I want to keep this article as simple as possible so as not to lose the point. I will follow this up with another article that investigates the conclusions of this article.

It turns out that in an environment where baserunners advance too easily because of errors, passed balls, wild pitches and stolen bases, the value of the extra base hit is almost neutralized.

My team lost a 4-run game and yet there were more similarities than difference in offensive stats between the two teams. This could lead one to question the value of statistics.

We had the same or very similar:

  1. AB (33 vs 34)
  2. Hits (15 vs 17)
  3. TB (21 vs 19)
  4. BB (10 vs 13)
  5. Errors (5 vs 5)
  6. WP (1 vs 1)
  7. K (3 vs 3)

The differences were:

  1. Singles (10 vs 16)
  2. OB (25 vs 30)
  3. SB (20 vs 7)
  4. Extra Bases (5 vs 1)

Looking at the above, we should have won. We had more steals and more extra base hits. Granted they got on base more often and 5 extra baserunners could have produced 4 runs given the 1.3 rate of baserunners to runs. But doesn’t power and advancing bases leave less men on base? Shouldn’t we have had a better rate of runs scored?

I first confirmed my intuition by using 3 different « Runs Created » formulas, plugging in all of the offensive stats including singles, doubles, and triples, walks, HBP, steals and caught stealing, sacrifices and grounding into double plays. These were the 3 formulas I used: Bill James’ original w/ Stolen Bases, Linear Weights version of 2002, and Pete Palmer’s Linear Weights in the Hidden Game of Baseball.

Each RC formula had my team winning by between 2 or 3 runs!

So what went wrong?

I took a look at the play-by-play. I immediately noticed 2 events not included in the summary statistics: Passed Balls and Outs on Base (Caught Stealing are counted but not Pickoffs or Outs trying to advance bases).

Passed Balls

My team did really badly here: we gave the opponent 18 extra bases by allowing 9 passed balls. They only allowed 2 passed balls giving us 4 extra bases. For some reason, PB is not part of the summary stats. Should be.

I added passed balls to the above RC formulas treating them like stolen bases. I added 18 PB + 7 SB to get 25 extra bases for the other team. My team only had 4 extra bases due to PB for a total of 24 extra bases (4PB+20SB). As can be seen, we are now equivalent in extra bases (24 vs 25).

However, passed balls have an advantage over stolen bases in that they result in more direct runs scored. We scored no runs directly as a result of our 20 SB (we never stole home). Whereas the other team not only stole home once, they also scored 5 times on the PB.

Outs on Base (OOB)

Here, we did only 1 better than our opponent: We got 2 OOB and they got 3.

By adding OOB to our RC formulas (by treating OOB like CS), our team gets an even better chance of winning.


Having evened out the number of advanced bases by adding Passed Balls to Stolen Bases, I thought I should also add errors into the equation. I broke errors into 2 categories: (a) the number of runners that advanced, adding this to SB and PB; and (b) the number who got on base because of an error, adding this to the on-base totals. The effect of all this manipulation is that the RC formulas finally had my team losing! but only by 2 not 4 runs…

Extra base hits

I then removed doubles and triples from the equation considering them equal to singles. I made the judgement (to be tested in my next article) that extra base hits are meaningless in the face of such a tide of extra-base baserunning (almost 30 for each team). So I removed extra-base hits from the totals and I finally got the exact score. My team lost by 4 runs both in reality and on statistical paper when I took into proper account the number of passed balls and errors, all outs on base, and removed the run expectancy benefits of extra base hits.