I just submitted this as an email to the SABR list:
Yaz had a career average against righties of .299 while it was .244 against lefties (that and all data mentioned here are from Retrosheet). So what do you think was his only season when he had a higher AVG against lefites? Yes, 1967, when he batted .338 against lefties and .323 vs. righties. In 1974 he batted .001 better against righties but every other year in his career saw him hit at least .030 better against righties. In both 1965 and 1966, he batted at least 100 points higher against righties. In 1968 and 1969, it was 46 and 79 points, respectively.
So it looks like among all the great things he did that year, he also ramped up his hitting against lefties, even relative to what he did against righties. His OBP was .027 better against righties, but that was the second smallest differential of his career (it .077 for his whole career). The differential was over 100 points in 1965, 1966, 1968, and 1969.
Now his SLG was "only" .500 against lefties that year while it was .657 vs. righties. That differential of .157 was even bigger than his career differential of .121
His OPS differential was .184, slightly lower than his career differential of .198. But, the differential was .391 & .27 the two years before 1967 and .205 & .332 in the following two years. So at that point in his career, his relative performance vs. lefties may have been unusually good.
I tried to calculate how many batting runs he had per plate appearance vs. lefties and vs. righties for each year in his career and then calculated the difference. I used the following run values
BB & HBP .33
Plate appearances were just AB + BB + HBP. His differential in 1967 was .0518, his smallest in 5 years and it would not be less than that for another 5 years. It was the 7th lowest of his career. The average differential of the two previous years and the following two years was .0997. Now that is .0479 higher than in 1967. Suppose he did .0997 worse against lefties in 1967 rather than .0518. Over 144 PAs, it would mean 6.9 runs. With 10 runs usually being equal to one win, it is possible that his much better than usual relative performance against lefties tipped the pennant to the Red Sox.
Added Point: He also hit especially great after August 31. He batted .417 (40 for 96). He ended August with a .3085 AVG. If he had hit that the rest of the way, he would have had 29.61 hits. So this created an additional 10 hits. He slugged .760 after Aug. 31. Through that date it was .594. This created an additional 16 bases. His OBP after Aug. 31 was .504 while it was .401. A rough estimate is that he had 2 more HRs, 1 more 2B, 3 more singles and about 11.6 fewer outs than expected based on what he had done prior to Sept. 1. Using the linear weights values, that generated about 7.89 more runs than expected. That is almost enough runs to generate 1 more win. But Sept. had the lowest league totals for AVG, OBP and SLG that year (probably not that unusual). YAZ greatly increased his performance while everyone else was going down. So it brings it even closer to one win and that is how many games they won the pennant by.