Sorry for the longish post but bear with me
I've been looking at the results in the QF, SF and Finals for the 100m.
There is an expected pattern - the runners run faster in the final than in the earlier rounds.
Compared with the semi-finals, the runners in the final are typically 0.05 seconds quicker on average. It might be a bit surprising that the difference is not more, but in almost all finals there are some who actually run slower than they did in the semi-finals.
There have been three occasions where the average time for the finalists in the final has been slower than their times in the semi-finals: 1952, 1956 and 1960. In the case of 1960 we know that one of the semi-finals was run with a 5 m/s tailwind, so its hardly surprising that the final was slower. In 1956 we know that the wind was swirling and difficult for the entire event so that is perhaps also unsurprising. Unfortunately there are no wind readings at all for 1952 so we don't know if that is a factor.
The top three average improvements from semi-final to final were:
3. 2004: 0.11 seconds. The fact that one of the semi-finals was run into a 1.6 m/s headwind and the final was run with a 0.6 m/s tailwind obviously affects the results here.
2. 1988: 0.13 seconds. Ignoring Johnson and Stewarts times. The final was run with a 1.0 m/s tailwind.
1. 1932: 0.23 seconds. This is the largest improvement in times from semi-final to final by a long way. The six finalists averaged 10.48 seconds in the final, compared to their semi-final times of 10.72 and their quarter-final times of 10.62.
This was the first year that wind readings were made, and they indicate a swirling wind. Three wind gauges were used and all showed quite different results*. Looking at the flags in the footage, the winds was obviously pretty strong outside the stadium** but that is not always a good indicator of the wind speed in the stadium. However, the times lead me to think that the wind was a factor in these races. Perhaps one (or both) of the semi finals had a strongish headwind and the final had a strongish tailwind.
Its interesting that in the 1936 final, which was run with a measured tailwind of 2.7 m/s, the average times were 10.50 compared to 10.54 for the same runners in the semi-finals, only a slight improvement.
* Here is a quote from an old post from DJ: "The IC4A report on the '32 Games talks about their various experiments with the gauges. Three were used for the 100m, the race in which both Eddie Tolan and Ralph Metcalfe ran 10.38/10.3. The gauges were placed at the start, the mid-way, and the finish, and gave respective readings of "+2.0mps crosswind," -1.4mps and +0.4mps."