Here's the rules on CN:
http://www.cyclingnews.com/road/2004/tour04/?id=stages/stage4
The official page lists the rules here:
http://www.letour.com/2004/us/index.html
A r t i c l e 2 4 - Team time trials
Teams start their team time trial at 5 minute intervals in the reverse order of the team ranking established following the 3rd stage. [...]
The stage ranking is established for each of the teams based on the finishing time of the 5th rider in each team (to the nearest hundredth of a second). An identical time is given to all teammates who arrive within this same time, even if they actually finish
quicker.
The procedure for establishing the general individual time ranking is as follows: Times for all riders are calculated according to the gap between them and the 5th man of the winning team (scratch time).
If this gap is less than that given in the table below, then the actual time recorded by the timekeepers will be taken.
If the gap is more than that given in the table below, then the rider will receive the time of the winning team plus the additional number of seconds indicated in the table below according to the place of their team.[...]
Riders who arrive on their own after the 5th rider in their team will be credited for the general ranking with the actual time they took to complete the stage.
However, riders finishing outside of the permitted finishing times (coefficient 6) will be disqualified.
As far as the general team ranking is concerned, the actual time of the 5th man in each team will be the time recorded for this ranking. [...]
The cited table is missing. However, CN claims it is 10 seconds per place to 15th, 5 seconds per place after that.
This needs serious work. Consider the following:
Team A: 70:00 (all riders A0-A8)
Team B: 75:00 (riders B0-7)
Team B: 75:01 (rider B8)
Team C: 80 minutes (all riders C0-C8)
Individual GC from this stage:
A0-A8: 70:00
B0-B7: @ :10
C0-C8: @ :20
B8 : @ 5:01
Huh? The guy finishes 1 second back from his team, and loses 5 minutes
as a result? Mind the gaps, guys.
This is an extreme example. However, it could easily occur that a 1 second
gap results in losses of more than a minute.
For 2005 an additional rule is needed -- no rider can lose more time relative to
his team time than the actual difference in time between the rider and the team.
But this is really too complicated. If they want to cap the time gaps, do
a numerical transform on the times. Calculate times the old fashion way.
The winning team gets its time. For the remaining teams, a transformed time
gap dt' is calculated from the actual time gap dt, as follows:
exp[2 dt/t0] - 1
dt' = -------------------- t0
exp[2 dt/t0] + 1
with t0 = 2.5 minutes, for example.
This would yield the following conversion:
dt dt'
(s) (s)
0 0
1 1
5 5
10 10
20 20
30 30
60 57
120 100
180 125
240 138
300 145
600 150
Too complicated? Hardly, compared to the existing rules. One step. Go faster, score better.
That's not so complicated.
Dan