Oftentimes, to win us to our harm,
The instruments of darkness tell us truths,
Win us with honest trifles, to betray’s
In deepest consequence.

—"Dave" Banquo in William "Author" Shakespeare's Macbeth

To win by strategy is no less the role of a general than to win by arms.

—Julius "Matos" Caesar

The high sentiments always win in the end, the leaders who offer blood, toil, tears and sweat always get more out of their followers than those who offer safety and a good time. When it comes to the pinch, human beings are heroic.

—George "Bamberger" Orwell, who obviously never read the ending to his book 1984.

Wesley: Excuse me? Is this something we should all be talking about?

Spike, Angel: No. (in unison)

—(Anaheim) Angel

Winning, that's the point of the game. Even a Phils fan like myself knows that.

So which are the teams that win? Well, obviously teams with more talent win more often. However, how do you measure that? Is it more due to their offensive prowess or pitching skills and how to quantify that?

A team's winning percentage in one-run ballgames is often cited by analysts (right after telling you a certain batter is 2-for-6 with a homer off a certain pitcher, as if that has any meaning). Are there certain traits, like winning close ballgames, that winning teams display? Can one predict a team's success from how they perform under certain conditions like close ballgames?

OK, enough with the question. After being directed to Baseball Reference's situations splits page by my friend Gregor Gross, and after kvelling an appropriate amount of time, I attempted to address these issues.

For offensive stats, I used the yearly team batting ratios: average, on-base, slugging, and OPS. For pitching, the following stats were recorded each year for each team: ERA, WHIP (walks plus hits per innings pitched), Strikeout-to-walk ratio (K:BB), Strikeouts-per-nine-innings (K/9IP). Then all stats were adjusted for by dividing by the league average. Finally, they were each compared against the team winning percentage and then the correlation coefficient, or how well each set of data corresponds to winning percentage, was determined for each decade and for all-time. I also adjusted the batting and pitching stats by the team's home ballpark factor and re-ran the correlation.

Then, I took the situational winning percentage for a set of criteria and tried to see how well each correlated to overall winning percentage. Those situations, which I kind of picked out of a hat, were the team's yearly record in home games, road games, one-run games, games won by no more than three runs ("Save" games), low-scoring games (five or less total runs), high-scoring games (10 or more total runs), and games with a large margin of victory (five or more runs). I also took the runs scored for and against in the above situations and calculated the expected winning percentage to determine if that correlated to overall winning percentage better (using the 1.83-exponent Pythagorean formula).

So without further ado, here are the results such as they are:

Battting

Of course, as baseball matured batting average became less important to winning, but don’t tell Joe Morgan. Overall OPS correlates best to winning percentage, but in the 2000s OBP is beating OPS, which was a bit of a surprise.

Park-adjusted Batting

When you adjust for ballpark OPS does beat OBP. However, you'll notice that none of these correlations is very strong.

Pitching

[Note: negative correlation just means that as one goes down the other rises, but still expresses a correlation. E.g., as ERAs go down winning percentages tend to go up. Makes sense.]

ERA correlates best here overall, though WHIP has surpassed ERA in the 2000s. The strikeout ratios are gaining strength but still don't correlate very well.

Park-Adjusted Pitching

ERA does surpass WHIP after the park adjustment, like OPS and OBP in batting.

Situational

Home and road records correlate equally well to winning percentage though for the last 25 years, road records have ruled. Also the expected home record beats out the road one. However, both of these are pretty strong correlations, which makes sense since they comprise about a half a season each.

One-run record don't correlate well and are doing worse as time goes by. "Save" records, close games, do correlate fairly well, which is why, I guess, they came up with the concept of a save in the first place.

This one's odd: Low-scoring record don't correlate very well at all, but high-scoring correlates extremely well, better in fact than based on the expected winning percentage. Maybe the Rockies should take note.

Again, it looks like higher scoring affairs are better predictors of a team's success or at least correlate better to their winning percentage. Perhaps this is due to a low-scoring game being more based on the pitchers' performance than the teams as a whole.