I have noticed that during raids many people’s Recount / Details differ quite significantly some times, and after the raid someone posts the relevant Warcraftlogs link which also seems to have its own opinion of what’s happened.
I presume everyone’s Recount is based on their own combat log, and it may be that due to lag or distance there might be differences?
Recount and details work the same but the reason they’re different is because recount starts recording when you get in combat and details when anyone close enough to register in your combatlog gets in combat. They both use the combatlog so range restrictions are there yes.
Warcraftlogs can get data from a couple sources, you can export your details combatlog for example or export a log created with /combatlog, the latter has a longer range than the in-game combatlog. Most common is the latter one with advanced combat logging in the network options enabled.
Recount does not take downtime into account while most other dps meters do.
Warcraftlogs sometimes doesn’t count padding on specific fights and/or has weird interactions on some bosses. For example on Skitra it doesn’t register the damage done by people that aren’t in the same phase as the person who is logging.
If warcraftlogs takes its info from one player’s combatlogs then it is also subject to all the failures/inconsistencies mentioned in my first post.
As far as I have determined Recount calculates DPS in a strange way but I was under the impression that Damage Done is absolute (unless there are range omissions I suppose). In other words look at Damage Done and ignore DPS which is somewhat arbitrary.
I wonder is there a way to merge two combatologs from two different players so as to get a more accurate log? For example if the timestamps are server based then perhaps this is not a difficult task, maybe a merge/union operation.
This is only a problem on skitra, but it can’t really be solved with the limitations and technology bliz allows access to. If we could see across the boundary to retrieve combat events, we could also make an addon that could do the skitra mechanic on its own.
Absolutely but to a lesser extent. The /combatlog command has a way bigger range than the normal combat log and is always recording so it’ll never miss any first hits for example but it is still one persons combatlog so it’ll never be 100% especially not on fights with different phases etc like Izilar mentions.
The reason is that recount uses active time to calculate the dps rather than elapsed time. Active time only records when you’re/someone* is attacking and not the time in between attacks so the damage will match details and skada but the dps will not.
As example I sometimes show myself as being top damage, my friend (healer) sees me as 12th!!!, and the 2nd on the list shows me as 2nd. Many other fights I see myself as 1 or 2, and then someone else’s warcraftlogs (usually the healer) shows me as 3rd, 4th and 5th even…
So either my recount fails to pick up other players’ damage or the healer’s fails to record (all) my damage.
And I cannot complain to the healer because I am not sure my Recount is doing it properly.
You basically shouldn’t look at warcraftlogs to gauge your own performance, especially at the moment.
People with higher percentile than you may either have better gear than you, better raid setup or simply better corruptions.
The best way to gauge your own performance is to compare how you perform compared to your expected simulated DPS while taking boss mechanics into account. If you consistently do between 90% to 100% of your sim DPS during a boss encounter (without taking adds into account as per a patchwerk sim) then you are already performing at the highest percentile level and are being held back only by your gear, corruptions, essences or sockets.
I for example am noted as 92nd percentile overall during an entire ny’alotha HC full clear, but I’m performing close to 100% and surpassed 100% during Maut (probably due to the ad spawn being cleaved) of my sim dps, which means I physically cannot perform any higher up to 100th percentile due to gear, essence and corruption disparity.
All in all, warcraftlogs is more used to see where you screwed up and how you died.