Understanding Meters and Logs (Absorbs)

[harky]

NOTICE: As I have retired from WoW, don't expect this to be cleaned up, or amended any time soon. Mods feel free to mess with it at your leisure if there is continued use for this thread. Extra note - this was written when the ICC buff was at 20%, all references to the ICC buff that do not also note a percentage of the buff should be noted as 20%.

First of all, what this is not:
- A guide on how to top meters. If you need one find a Druid healing guide.
- An argument for, or against meters. If you don't think meters are important then you're hurting yourself and not me.
- QQ. If you want a thread to whine about 'losing' on meters find another one. This game is not perfectly balanced, get over it.

What is it? Well, hopefully it will help dispel a few myths about meters and demystify the problems with meters and absorbs. Meters are only as good as the person interpreting them. Not understanding them leads to a lot of misleading information. There are two things I'd like to go over:

- Absorbs
- Absorb stealing

Warning: This thread is TL;DR! If you'd like to comment that it's TL;DR, just stop reading now. Thanks.

Absorbs:
The first thing to keep in mind on absorbs is the most important. No meter records absorbs properly. Meter's accuracy of absorbs depends mostly on the other information provided by the meter and how it can be used to get a proper picture of how much damage your spells absorbed. If a meter provides thorough information it can be used to properly evaluate shielding, if it does not then it should be dismissed and not used for checking absorb performance. For this reason the four common meters should be ranked as follows in regard to absorbs: WoL > WMO > Recount > Skada. WoL has the most comprehensive breakdown and will allow you to weed out inaccuracies the best. WMO is a very close second. Recount is still solid, but it does not provide the depth of information needed. Skada is by far the worst and I do not under any circumstance advise you to use Skada to track absorbs.

Related to that issue is also the combining of healing and absorbs. If a meter is not sufficiently comprehensive healing and absorbs should absolutely not be combined. This is another mark against Skada, which is most often used in Healing + Absorb mode. Recount is decent about this as it typically keeps Absorbs separate. WMO and WoL are both detailed enough that they can combine the two safely. Note that Skada and Recount 'typical usage' is not taken into consideration when ranking them. Skada is inferior to Recount for other problems with reading absorbs. The reason for keeping the logs separate is because of corruption. The healing done is extremely accurate, but remember that first rule: No meter records absorbs properly. This means you are combining an inaccurate meter (absorbs) with an accurate meter (healing). While healing is very accurate, when combined the two become untrustworthy. Very complex meters, like WoL and WMO have the two listed together, but also tell you exactly where the healing and absorbs are credited, how they're credited in what percentage they're credited and so on. The information provided by Recount and especially Skada is not by any means in depth enough to allow this to be combined. As such in-game Absorb + Healing meters should be avoided at all costs. If you must post meters in game, for some ridiculous reason, always post healing and absorbs separately to avoid polluting the healing charts.

Once you understand that the meters can not be taken at face value you need to know how to interpret them. The current best way to do this is spell counting. What you need to keep track of is the number of PW:S glyph hits and the total amount of crit healing. From this point all you need is some simple formulas.

PW:S formula (Discipline only): ( 2230 + 1.2068 * Spell Power ) * 1.04 * 1.05 * 1.15 * ( 1.2 for ICC w/ buff only)
Divine Aegis formula: Total critical healing * .3

There are two ways to do this last part, one is simple and in general should be what you do. The other is more complex and should be really be reserved for if you find something else fishy. The first is as follows:

Step 1: Calculate Divine Aegis maximum absorb.
Step 2: Subtract the maximum DA absorb amount from the displayed DA absorb amount.
Step 3: Add the remainder to the displayed PW:S value.

Now there is one other problem that requires some guess work, but does lead to some issues. DA's absorb amount includes over-healing. This means that your total critical healing as displayed will typically not include over-healing, but your DA absorb amount will.

I'm going to use a random Disc Priest on Blood Queen normal 25 for this, which can be seen here: http://www.worldoflogs.com/reports/2...?s=2627&e=2933

Total effective critical healing: 277,986
Total DA absorb: 1,284,161

This DA recording would require a total healing of 4,280,537. As such this DA value is wrong. Now we need to calculate how much total critical healing was done. This is partially guesswork.

Step 1: Select a healing source and record its total healing.
Step 2: Divide that number by the non-over-healing percentage.
Step 3: Multiply the result by the critical amount% (highlight the heal value to see this number)
Step 4: Return to step 1 until all healing sources are recorded, then proceed to step 5.
Step 5: Add all the step 3 results together.

So let's do this for that logs glyph of PW:S healing to start.
Healing = 316,980
Over-heal% = .443 (44.3%)
Non-over-heal% = 1 - .443 = .557 (55.7%)
Total healing = 316,980 / .557 = 569,084
Critical amount% = .502 (50.2%)
~Total crit healing = 569,084 * .502 = 285,680

Repeating this for the other healing sources and we wind up with a total approximate critical healing amount of 537,
642. Please keep in mind that this is not a completely accurate number. I actually like this log precisely for this reason. You can see in this log a 100% over-heal Prayer of Healing, 80% of which crit. Without going through the log line by line we can't know how much this would add. However, it is a good rough idea.

Now all we have to do is take our 537,642 total critical healing and figure DA, which is very simple: 537,642 * .3 = 161,292. This number is the actual amount of absorbs provided by Divine Aegis. Next we take that and subtract it from the original DA absorb figure of 1,284,161, which gives us: 1,284,161 – 161,292 = 1,122,689. This number is falsely attributed absorbs. In general this may be added directly to the PW:S number. The PW:S number in this log is 1,744,161. Combined that works out to 2,866,850. This is what you would call the 'real' PW:S absorb amount. This means that PW:S was not 48.7% of healing, it was 80% of healing. It also means that DA was not 35.8% of healing, but 4.5% of healing.

Hopefully this will illustrate two things. First the complete inaccuracy of the meters, but second that with proper understanding and proper data you can actually get coherent data out of absorb meters. You will never get a completely correct number without doing cast by cast analysis, but we can take a meter that gives a better picture of where absorbs are coming from.


Absorb Stealing:
Yes, there are ninja super pirates in your raid that will not only steal your absorbs, but they will use them for their own nefarious reasons to corrupt and polite your otherwise rapture filled experience. In short they will borrow your time. It takes great mental strength and agility to overcome their twisted faith. However, with focused power and will you can bring grace and enlightenment back through meditation to ward your soul, suppress your pain and make good on the aspirations of all Priestly goodness...

Okay all joking aside you've probably never heard of absorb stealing. Believe it or not, we just went over it a minute ago. Remember how we had to figure out how much of PW:S's absorbs were credited to Divine Aegis? Well, that's absorb stealing. What happens is because the combat log absolutely does not support proper absorb reading a meter must make assumptions. Key among these is what did huh to which now? I digress. It's actually very simple. The last ability that causes absorbs on a target will be credited all absorbs until that spell fades at which point they will be credited to the next most recently cast spell and so on.

So if you cast PW:S glyphed and the glyph heal crits it will trigger Divine Aegis. That's a good thing. However, PW:S is applied first. This means that whenever you crit with the PW:S glyph heal this happens:
1.PW:S bubble applied.
2.Glyph heal crits.
3.Divine Aegis is applied.

Problem? Well, that's pretty easy: The Divine Aegis absorb amount triggered here will be less than 10% of the size of the PW:S cast. However, because Divine Aegis was the last absorb mechanic cast on the target it means that for the next 15 seconds any damage absorbed will be credited to Divine Aegis until the absorbs break entirely. So let's say you shield three people. One glyph crits, the other two don't. All are fully absorbed. Now, what does your meter say? Let's assume you shield for 10k, heal for 2k, crits for 3k. Easy numbers. So should have 30k absorb for PW:S, .9k absorb for DA, 7k healing. The meter will instead claim 20k absorb for PW:S, 10.9k for DA and 7k healing. Divine Aegis 'stole' your PW:S.

Not a big deal, right? You 'healed' 37.9k damage either way.

The problem is this: You are not the only one in the raid who can use absorbs.

Who can use absorbs that clash with yours? Other Disc Priests using Divine Aegis. Anyone with Val'anyr. Holy Paladins using Sacred Shield. Mages using any of their absorbs. Warlocks using Shadow Ward. DK using Anti-Magic Shell.

Basically anything except PW:S. However, even that isn't safe. If you have PW:S your Weakened Soul (or stronger shield) will protect it from being overwritten. On the other hand if a heal you cast crits on someone and another Priest then puts PW:S on that target they will then get credit for your Divine Aegis proc.

This is the same reasoning behind the old Twin Valkyr log bugs. Throughout the fight you are absorbing one type of damage. For the entire fight. Like, all of it. Meters do not figure in this damage to meters. However, as soon as you put another absorb on someone all damage absorbed is credited to you. So you're taking constant hits of 3k damage which are being absorbed and another hit of 3k damage which is hurting you. If you then shield yourself the meters will credit you for 6k absorbs for each hit even though you're only preventing 3k damage. This is why logs for Twin Valkyr were basically impossible to read for most. It was fairly simple though really: Divide everyone's absorbs in half. That gave a pretty decent reading. It's also why many Druids were completely unbeatable on that fight with Val'anyr and posted some outrageous logs. It's also why HPS on Valkyr is essentially a joke and was purged from major log sites a few times.

What you need to check for to evaluate things is looking at the healing done by classes that commonly use absorbs. This will be anyone with Val'anyr, Paladins, Death Knights and other Discipline Priests. Though if you have two in your raid there are other problems to be addressed. So, how do you spot the problem?

Other Discipline Priests: If you think something is fishy check their maximum DA, then their PW:S. Now if it still seems odd divide the total PW:S you came up with by the total number of hits from the PW:S glyph. Does the result seem rational? I'll go back to the previous log I was using which we figured was 2,866,850 absorbed. He has 241 PW:S glyph hits. 2,866,850 / 241 = 11,895. Now is that in ICC? Yes it is. This means we divide that number by 1.2 for 9,913 per shield normally. Now how much SP would that take? Just reverse engineer the formula: 9,913 / 1.15 / 1.05 / 1.04 – 2,230 = 5,663. That's the amount from SP. Now divide by 1.2068 and we get 4,693 SP reqired. This is reasonable and we can assume it's accurate. If it were not believable we would check the players armory and see if the SP figure was remotely possible. 4.7k raid buffed definitely is possible, so there's no need in this case.

Death Knights: This typically isn't a big problem. Check their AMS casts and compare to their health pool. It should be 50% of health multiplied by the number of buffs gained and then multiplied by 1.2 if in ICC. If the total isn't pretty far off, it's possible that they stole some absorbs.

Paladins: This is a little harder. What you can do is check who their Sacred Shield was on, then check how often you used PW:S on that target. If there's a lot of overlap you were both most likely stealing substantially. Typically I would call this even and say just try not to shield tanks so often unless you're tank healing. If you are tank healing and a Paladin is healing that same target good luck. I know of no way to accurately figure that out. Sacred Shield will proc and steal DA, DA will steal Sacred Shield and both will steal PW:S. It's horrible.

Val'anyr: This one isn't as hard as you might think. The key here is to know how much Val'anyr can actually absorb. If there's only one in your raid it's pretty easy to figure this out. Especially if you're the only Discipline Priest. Val'anyr's buff works out to approximately 5% healing. It has ~1/3 up-time and for the duration it will produce 15% of healing in absorbs. Thus roughly 5%. If you ever see someone being attributed much more than this throw up a red flag. Quickly take their healing and multiply by .05, then subtract that from the Val'anyr amount. Only do this is the result is very far off, don't go nuts over anything less than about 8%. It's technically possible that a player was on a roll during the proc and their HPS spiked producing an abnormal amount of healing during the Val'anyr buff. However, if you see it up over 8% immediately take the extra healing and reduce that to 5%, take the remainder and set it aside. Now compare PW:S/DA figures within the raid and check as you did above for verifying PW:S numbers against another Discipline Priest. Do those numbers make sense? If they do not line up try attributing the extra absorbs you set aside to PW:S. Does the additional amount make the numbers closer to the expected values? If so it's very likely that the Blessing of Ancient Kings buff stole from either DA, or PW:S. It's not possible to know which, but because we isolated DA we know that figure is very close to being correct, which leaves the PW:S amount in debate. See Addendum #1 for further information.


The final thing to keep in mind and I can't say this enough: Placement on a meter does not matter. Meters should be used to evaluate performance, which should not be done in a competitive fashion. When trying to analyze your performance or the performance of another healer keep it objective.

[Calendra12345]

great post, thanks

[Danner]

A good read, Harky.
Thanks!

[Myzrael]

Well written and easy to understand, helped me get "enlightened", thanks.

[Lysdexic]

harky for president

I always love looking at WoL and seeing roughly 50% of a person's absorbs being put as DA.

Yeah so I was bubble spamming most of this fight and my absorbs are like: 750k PWS and 700k DA. DA is really op for a 3 point talent!!!

[Kelesti]

harky for president

I always love looking at WoL and seeing roughly 50% of a person's absorbs being put as DA.

Yeah so I was bubble spamming most of this fight and my absorbs are like: 750k PWS and 700k DA. DA is really op for a 3 point talent!!!

It was even funnier when one log of mine had Circle of Healing giving me 18%, and Divine Aegis giving me 21%. I didn't know I could have both.

[harky]

Addendum #1: Val'anyr

This is for clarity sake. When referring to the 5% healing from Val'anyr, that is 5% net healing. To calculate net healing you need to incorporate over-healing. To do this you must take all spells that over-healed and increase proportionate to their over-healing figure.

As an example I will use the following log: http://www.worldoflogs.com/reports/a...=11360&e=11666

This is simply a log found at random with Val'anyr being used with an 'abnormal' healing figure at 10.7%. There is a complex and a simple way to do this. Complex would involve doing each spell one by one, however I find that unneeded. Instead I suggest taking total healing and dividing by the total non-over-heal percentage. You find this the same way a you did with Divine Aegis. This total over-heal percentage is 31.2%, so the non-over-heal percentage is 1 - .312 = .688 (68.8%). Total healing is 5,006,386, so 5,006,386 / .688 = 7,276,723. Now, how much was the Val'anyr healing? 533,510. Now to find this percentage simply take that number unmodified and divide by the net healing number: 533,510 / 7,276,723 = .073, or 7.3%. So this number is very close to rational.

There is technically about 170k more there than expected, so let's compare the Discipline Priest involved, which is here: http://www.worldoflogs.com/reports/a...=11360&e=11666

DA comes to a real figure of 127,490, with 342,271 falsely attributed from PW:S. This is a total adjusted value of 1,166,535 for PW:S. This is over 140 casts so the average cast absorb 8,332. This already looks very low. Low enough to throw up an alarm; so I'll check his SP on armory... Unfortunately said played is logged out in silly joke gear. Goodie! So we'll continue on to see how much SP this would be. Looking really quickly we find Hellscream's at 15%. So we take 8332 / 1.15 / 1.15 / 1.05 / 1.04 - 2230 = 3,539.4. Now the scaling value: 3,539.4 / 1.2068 = 2,932.88. Rounding up to 2,933.

Is this number rational? I'm assuming no. This would require the player to be in Naxx 25 and Ulduar 10 gear, with perhaps some other crumby gear tossed in. While I can not see her gear, she does have some 277 pieces. As such I am going to credit that 170k from Val'anyr to her instead. This increases her PW:S number to 1,335,535. This now brings the average per cast up to 9,539. Does this sound more rational? I would say yes. Let's reverse engineer again: 9,539 / 1.15 / 1.15 / 1.05 / 1.04 - 2230 = 4,375. Adjusted for scaling: 4,375 / 1.2068 = 3,625.

In this case the characters armory is no help, but all you really need to ask is this: What is more likely, a Disc Priest with hardmode kills with 2,933 SP, or 3,625 SP? To me that's a very easy call.

Now this may seem to contradict my previous statement calling Val'anyr 'easier than you think', but keep in mind that almost all the work here was actually on the Priest, not the Druid with Val'anyr. All we did with Val'anyr is adjust his healing by his over-healing, then compare the percentage from Val'anyr. However, it is important to understand that Val'anyr, like Divine Aegis, does include over-healing. Luckily it's also much easier to factor.

PS: WTF sticky?

[Waterwings]

harky for president

I always love looking at WoL and seeing roughly 50% of a person's absorbs being put as DA.

Yeah so I was bubble spamming most of this fight and my absorbs are like: 750k PWS and 700k DA. DA is really op for a 3 point talent!!!

I've seen logs where DA was above Pw:S; YAY

ot:Good post harky!

[Lysdexic]

I think this post should be written on harky's tombstone. It will really bring out his character. RIP.

Edit: The opening post of this thread...

[Deleted]

I think this post should be written on harky's tombstone. It will really bring out his character. RIP.

Did I miss something?

[fabian]

Did I miss something?

The first sentence of the OP? He's retired from WoW.

[Lysdexic]

Ok I think you guys are missing the scope of my trolling. Just scroll very slowly, from the top of harky's post to the bottom. Tell me how long it takes you to scroll the distance. I was feeling nostalgic for ridiculously long walls of text. Having this post, just one of his many, on his tombstone would give people a good idea of what kind of person he was.

[Lysdexic]

Can this get renamed to something like... Understanding Meters (Absorbs)
orrr
Guide on Absorbs & Meters
orrr
well you get the idea

[Kelesti]

Done.

[Marraphy]

This just confused me ^^;;;

It would be great if they could tweak the meter addons to do this for you...