I always seemed to have the impression that Degrees of Success in opposed tests were calculated reducing the results (test DoS = winner DoS - loser DoS, if any), in case both parties succeded.

Reading the rules for Only War, my understanding seems to be that DoS in opposed tests is winner's pure DoS, and the losing side succeding in the test does not reduce the result in any way.

I don't want to talk about right or wrong, but is that understanding shared by everyone reading the rues? And also, was it always like this in previous games?

rafaelmb said:

I don't think it works like that, here is the section labled Opposed Tests on page 31.

"Ther are times when a character needs to Test himself against a foe, and times like thse call for an Opposed Test.  In an Opposed Test, each participant makes his appropriate Test normaly and who ever succeeds wins the test.  If both parties succeed, the party with the most Degrees of Success wins the Test.  If the number of success in equal, then the party with the hightest Characteristic bonus wins.  If the result is still a tie, then the lowest die roll wins.  Should both parties fail, then one of two things occures: either the Tests ends in a stalemate and nothing happens, or both parties re-roll untill there is a clear winner. Either outcome is appropriate and is left to the GM's discretion."

The winner of an opposed roll counts all his degrees of success if it matters beyond the opposed roll itself.  If someone tries to sneak past someone else and they succeed with 3 degrees of success and the person trying to notice them gets one degree of success then they successfully sneak past them and count all 3 degrees of success for any other reason.  In that same scenario, if the sneaker has attempted to get past multiple opponents then they would all roll against the sneakers one roll and if any of them had a better result they would spot them, but each successive roll would not reduce the effectiveness of the original sneak attempt.

Another example would be two people running a race.  One persons speed does not slow down the other person, so why would you subtract one persons result from the other?

Yupsate said:

That technically isn't an opposed test though, for chases (and races) you normally both try to get a certain number of successes before the other character. A better analogy for an opposed test is a tug o war or arm wrestling contest.

I've always played it that the losers degrees of success (if any) do reduce the winners DoS since the early days 40K rpg's but the rules aren't clear. I don't think it makes that much difference as normally success is enough on opposed tests but may start to make a large difference later on when you have a Force weapons and the like so I'd advise you to use the 'rules question' link at the very bottom of the page.