I was thinking.
It would be silly to refresh the multiplier during the round ?
In current system it is useless because it bothers only with victory or loss.
But if we take kill- and surviverate into account.
Updating during the round has two basic problems:
1. SURVIVALRATE will be very high for a long period, then drop sharply as the main forces clash, then stay very low while trying to corner the last archer. KILLINGRATE will follow a similar (reversed) pattern. It will make the multiplier go to extremes. These statistics make sense only on a scale of a full round.
2. Who is the winner and who is the looser (apply SURVIVALRATE or KILLINGRATE)?
In principle, we can think of a way to weigh the two statistics and get a net bonus (one can only increase and the other only decrease). Then it could make sense. The good thing about such a dynamic mechanism is that when the situation is hopeless, instead of dragging time and getting hit by more unfavorable "ticks" to the multiplier, the best strategy would be to stage a last suicidal charge to take down as many enemy with you as possible.
But it might just complicate things, so better leave it a round update.