There's an extensive literature on the statistical analysis of baseball strategy.
One technique is to apply multivariate regression to determine the number of runs a team expects to score as a linear function of their singles, doubles, walks, and so forth. One can then use this to assign a run value to each individual player based on the individual accumulations of singles, doubles, walks, and so forth. This achieves a fairer and more reliable estimation of player value than the traditional measures of individual hitting value, such as runs scored or batted in.
Other methods allow one to discover that some baseball parks strongly favor the pitcher, while others strongly favor the batter, and to adjust the estimates of player skill to compensate for this. For example, Greg Maddux pitched half his 1993 games in Fulton County Stadium, a notorious hitter's park. By comparing his home pitching with his road pitching, and by comparing his pitching in Fulton County with other pitchers', we can get some understanding of how well he would have pitched in a more neutral park.
Similarly, one can use statistical methods to compare players of different eras who may never have played at the same time.