Can you give an example calculation to show that this would actually have any effect?
It is a weighted average, currently:
FFR = sum(rate * amount) sum(amount)
and I'm proposing:
FFR = sum(rate * amount * age) sum(amount * age)
age is the age of the open swap, in an arbitrary time unit (eg: hours), starting from a moving epoch (eg: 30 days ago).
The end effect will be a more dynamic FFR rate, that adapt to swap market conditions faster (but it is also less stable).
A nice side effect
Right now opening a swap from the FFR offers cluster
can't mathematically increase the FFR rate itself,
so the cluster need to be consumed together with higher-rate offers to cause even a minimal increase.
After weighting by age, however, opening a swap from the FFR offers cluster
will mathematically increase
the FFR rate (at the next update, eg: at the start of the next hour).
Additionally, weighting by age, the FFR will change even without new swaps being opened or open ones being closed.
This can have some utility too, in the rate-discovery process.