Well, I wasn't involved with the 1996 Atlanta Olympics so I don't know what they did but I believe that the precision of measurement is probably better now than it was then, and that precision means we were able to do certain things that I think would be hard to do with less precision.
Regarding the drafting giving a boost the the front rider, a while back I saw an examination of this that looked reliable. Sorry, no link, but while there could be a boost in theory it's tiny, and to even get that tiny boost the drafter would have to be practically glued to the cyclist's back. Millimeters. So for all practical purposes, as far as I'm concerned it's busted.
It seems petty simple to me; there's some theoretical reason to believe that the front rider gets a very small benefit, but at normal bike speeds there may be no, or virtually no effect.
There's no robust data to support the existence of an advantage to the front rider. There is some limited data from computer simulations, and less than rigorous observations to support that there may be a tiny effect. So it's still open for further study.
And the fact that it only matters to determine who wins a pedantic debate on a bike forum may explain the lack of conclusive, rigorous testing to resolve the issue.
You could fall off a cliff and die.
You could get lost and die.
You could hit a tree and die.
OR YOU COULD STAY HOME AND FALL OFF THE COUCH AND DIE.
From what I've heard the slipstream only projects back about 6 feet. This means that the third rider is not seeing a benefit from the slipstream from the first rider.
Here's the link to the published article (which few can read because you need a subscription, hence the preprint):
Remember that you're looking for changes in the overall drag, so a 10% error in the model really means that 3% change is drag would be 3%+-0.3%.
The Blocken paper seems to be independent authorship from the web link, so you've got two groups basically in agreement, compared with no credible disagreement.
A link I ran across months ago (can't find it now) indicated that the third rider benefits more than the second.
Just throwing some numbers into analyticcycling using myself as a model I get that a 1% decrease in Cd (with in the range of the Blocken and Ferguson findings) would reduce the watt requirement by about 2.5 from 300 to maintain 27mph.
Personally I consider that to be most likely not noticeable or meaningful in a rotating pace line. But since multiple sources indicate that a very small gain is probably realized under the right conditions while cycling, why so interested in fighting about it?
At something around 1% that is going to be tough. Figures from the Blocken study show the gap length and riding position of the riders having many times the effect of a possible 'push'.
I always hate getting behind the 110 lb, 5 ft tall little guy that rides in his drops and barely breaks the wind at all. Its almost like riding up front. Now getting behind the big behemoth is nice bonus and will make a big difference in effort and can add up on a long ride. There's so many variable to consider and the size of the rider and how close you tuck in are just a few.
So what would happen if we got enough riders together to circle a track, we are all in the draft, how fast?
The amount of benefit varies with the number of riders but also their speed. For a single rider on a 250m indoor velodrome at individual pursuit-like speeds the reduction in "effective" drag area due to this effect can be close to .005 m^2.
Field measurements of aerodynamic drag based on velodrome tests need to be corrected for this effect. That is, real-time measurements of drag will decrease for the first few minutes while the rider "stirs the bath tub" until it reaches an equilibrium value. You can also see this in the power-to-speed relationship for hour record attempts and in team pursuit events. If you do field tests outdoors in calm conditions, you can see the effect on aero drag if another rider is even 15 seconds or so up the road (if a car passes, you can see the effect for at least 30 seconds).
So, I don't know where the "6 feet of slipstream" came from but that sure ain't what I've seen.