We all feel that 1st govs in preliminary rounds are less successful than all the others, but we also see that 1st govs win finals fairly often - in Germany between DDM Berlin 2008 and ZEIT DEBATTE Stuttgart 2010 at almost every single tournament. Does that mean that good teams perform better in 1st gov?
Here is a look at the EUDC 2006 tab: 164 teams provide after 6 rounds 984 results.
We can see that 1st gov's odds to gain three points are - as expected - generally below average. Betting on a 2nd opp win seems to be a way better idea. Now, looking at the top 50 teams only, we see the respective percentage is above average. But top teams are much more likely to win in any position. Taking this into account, we learn: chances to gain three points are still the higher the later you speak in a debate, and the best way to avoid a loss is being some sort of opp. That goes for the totality of teams as well.
However, this analysis does not cover break rounds. Possibly, dynamics are different before an audience in everything-or-nothing mode.