That is indeed the reason why that happens. Healing does not reset the fight timeout counter. We chose to do this because you will often heal up after a fight, which should not count toward the fight time.Killshot wrote:Our Iron Juggernaut logs are getting split into weird parts:
We don't attack the boss during Siege Mode at all (siege mode lasts 1 min). That may be the reason why this happens. There's a lot of healing during that phase though. Shouldn't healing and damage from the boss reset timeout?
Killshot wrote:I don't see a single reason why Seismic Activity can't extend combat. Boss has GUID and you can easily verify it's same boss damaging you.
Does Seismic Activity always cause players to take damage? Buff gain/fade events are notoriously unreliable, so I'm hesitant to track & use those for time segment extension, but damage events from a given source NPC and spell ID combination are definitely a way to do this.Kihra wrote:Yeah, that would work. Requires special case Iron Juggernaut code, but should work. If nobody is taking damage, then you'd have to track him gaining the buff and stay in combat until the buff fades. If he despawns on a wipe, though, it might not fade.
I'm going to do three things to try and resolve this. First off I think our servers currently still have a special case for Immerseus that makes them consider damage events with overkill to be death events, causing all these fake kills. This has worked very well for other encounters in the past, but damage events with overkill seem to be generated every time Immerseus submerges. I'm going to disable this special case and rely solely on actual death events for now.Liquidsteel wrote:Immerseus heroic wipes being logged as kills. I know you're aware of this, just posting my logs to see if it helps. We had 10ish wipes before the kill, but half of our attempts are marked as kills, inflating ranks.
Secondly, I will add a different special case that will consider events where his reaction turns friendly to be death events. This may resolve issues where an actual death event is absent.
Finally, I will enable validation based on the number of killed & purified puddles, with a threshold to be determined, probably 90 corruption. That should filter out most other outliers.
Yes and no. I agree that flex raids should not be ranked as normal. However, I do not see the relevance of rankings different types of flex raids against eachother. In my opinion, flex raids are too diverse to make for very interesting statistics when comparing them to one another. Work is in progress to extract boss HP values from the combat log fields added in the 5.2 patch and use those to detect the boss difficulty. This may allow us to separate flex from normal and not rank them at all. Suggestions are of course welcome.Liquidsteel wrote:Also, are there any plans to separate flex and normal ranks?