That wasn't really a leftover of this commit.
I didn't do anything here at the time as doing something here would have other (design) implications which needed discussions.

testAuras("global", (name, cmpAuras) => {
TS_ASSERT_EQUALS(ApplyValueModificationsToTemplate("Component/Value", 5, playerID[1], template), 15);
playerState = "defeated";
cmpAuras.OnPlayerDefeated();
TS_ASSERT_EQUALS(ApplyValueModificationsToTemplate("Component/Value", 5, playerID[1], template), 5);
}

could have been better

Which use case could apply to applying auras to defeated players?
Isn't it a clear design if defeated players count as non-existant players?

To be continued at D1430.

Remark was mainly about gaia.
But anyway, the suggested optitmization is not related to this specific commit.

It said "Remove team bonus when the source player is defeated", so I commented it here as to me it looks like this commit should have therefore introduced a defeated check in this line. But blame is not so important as getting it fixed. Do you agree that I can add the defeated condition here?

I still don't see the link with that commit ('source' vs 'target').

I have not to agree with anything right now. But removing defeated player when recomputing affected players seems probably wanted and probably an optimization. Though it's hardly noticeable. (All target entities but player entities should already have been caught somewhere when changing ownership to gaia, so the only remaining case is when the player entity is the target. Example : resource counter of a defeated player which continue to increase). I didn't/don't know what is done for techs.