In some undefined edge case, honorspy re-set the estimated honor to the actual blizzard "this week" value
This can lead to large difference in real honor if it happens in the middle of a PVP night.

I have a specific case I can describe here that will probably not help solving it but should help understanding what I'm explaining.
I had a player aiming to stack his honor that was going to take a plagueland tower : He had 106634 honor and had few HK during the day.
When he gained the tower honor, his honor felt back to 106299 (blizz week value + tower honor, HK honor was lost)

Here is a screen of his character sheet after the tower

I can't find out how and why yet, still investigating

I've been able to debug a bit, here is what I found.

At HonorSpy:CheckNeedReset I had this print:

    local _, thisWeekHonor = GetPVPThisWeekStats()
    if (HonorSpy.db.char.original_honor ~= thisWeekHonor) or (HonorSpy.db.char.last_reset ~= must_reset_on) then
		print("HonorSpy:CheckNeedReset", "reset", skipUpdate, HonorSpy.db.char.estimated_honor)
		print(HonorSpy.db.char.original_honor, thisWeekHonor, HonorSpy.db.char.last_reset, must_reset_on)

And when it happened to me the output was (line 2 and 3)

[16:51:56] HonorSpy:CheckNeedReset skipUpdate nil 80759
[16:51:56] HonorSpy:CheckNeedReset reset nil 80759
[16:51:56] 49218 0 1656486000 1656486000
[16:51:56] HonorSpy:CheckNeedReset2 reset nil 0
[16:51:56] HonorSpy:CheckNeedReset skipUpdate nil 0

The player thisWeekHonor was 0 at the moment. Could the GetPVPThisWeekStats return a wrong value in some edge cases ?
The point is I also had this debug print :

function HonorSpy:CheckNeedReset(skipUpdate)

	print("HonorSpy:CheckNeedReset", "skipUpdate", skipUpdate, HonorSpy.db.char.estimated_honor)

	if (not skipUpdate) then
		HonorSpy:UpdatePlayerData(function() HonorSpy:CheckNeedReset(true) end)
	end

And as you can see on the output above, HonorSpy:CheckNeedReset is only called once with skipupdate true, without a skipupdate nil first and I can't find any other call to HonorSpy:CheckNeedReset(true) in the code outside HonorSpy:CheckNeedReset function.

If GetPVPThisWeekStats is the culprit here, we can't control the faulty behavior.
In this case I'm considering a timed call back (5-10s ?) to checkneedreset with a "doublecheck" boolean when the returned value is 0 so it does confirm a second time the actual value and doesn't reset on the first missreturn

Any opinion about this workaround ?

We can try that, maybe you can test it for some time locally first and see if GetPVPThisWeekStats is the actual cause?

I'm definitely debugging more on my side before even putting up a merge request yeah
I'll keep you updated

I unable to reproduce this issue for a long time now. It was quite common when I first opened this thread.
My fix should be functionnal and ready but I didn't had the chance to test it properly because of that
Maybe something changed on blizzard end or in the context somehow ?

I will keep my debug/test local version.
Maybe I can open a draft merge request with my WIP fix and we let it be so it's there if it's ever needed later on ?

Sure, let's keep it for history

Happened to a guildmate lately, so it's not fully gone