Raid Rx: Analyzing you healers Part 3
Raid Rx is designed to encapsulate and cure the shock and horror that is 25-man raid healing. Ok, so it's mostly horror... Anyways, if you're a big fan of X-TREME Whack-A-Mole (or are being forced into it against your will) this is the column for you. A lot of times when I'm sifting through data on WWS, I feel like I need Sherlock Holmes' hat and magnifying glass. Today is no exception.
If you're just joining the series for the first time, here's a link to get you caught up. The rest of you should recall that we left things on the cusp of actually going through WWS. Today we will rectify that and get into the nuts and bolts of WWS healing analysis. See you healing sleuths after the break!
First things first... If you're going to talk about WWS, you should have some example data to go along with it. I give you Anonymous Report, which will be good until 2 weeks from today. Otherwise, feel free to follow along with the posted pictures. Each one is a link to a larger version for easy eye balling. Also, keep comments on-topic. We're not here to compare the length of our dps or how much cooler your tanks are.
The first thing you run across by clicking the Heals tab is the total healing for the duration of the report. In this case, it is for the first 7 bosses in Black Temple, including various wipes. Here's the background info you need to know before you start analyzing.
The first 9 healers are healers. The paladin, Gemini, left after the second boss. Rain and Aquarius are new to the guild, working to gear up from T4/T5 content. They're not familiar with the bosses in BT and don't have the SR gear to do Mother. Calliope is a shadow priest. Taurus came in for the last 2 bosses. Polyhymnia is an imp spi priest. Typically druids, priests, and paladins are assigned to tanks. Shamans are on raid healing. But we'll break those down by fight.
So what do you actually get out of looking at total healing? Not a whole heck of a lot when you have people swapping in and out. Resist the urge to compare one class to another. Part 1 will give you a recap if you forget why it's a bad thing.
That pretty much leaves us with comparing Hwesta to Harma. The rule of thumb here is if people of the same class have about the same assignments, gear, talents, and experience, and they're within 5% total healing of each other, they're doing fine. That 5% normally absorbs random things like afks for cats on fire, bio breaks, and phone calls. If I see the difference approaching 10%, I'll immediately check out their spell selection and hits/ticks.
Overhealing is actually complex enough to have its own post, which you can find here. There you'll find much more in-depth info on what's good and bad OH. For our exercise here with T6 content, 40-60% doesn't really phase me at all since it's the cost of being proactive.
So let's pretend one Shaman was significantly lower than the other. The easiest way to compare players of the same class is to click the drop down menu next to Browse and select Raids & Mobs. From there it's as just pick out the class you want to compare. For us, this means clicking on "5 Shaman". The table that pops up contains every spell all 5 shamans used or experienced during the entire evening. Scroll down to the blue healing section. Under Harma for Chain Heal you're going to see a light grey 2442 and 14%. This color indicates are the average amount per cast and the chance to crit. The light purple 2991 is the number of times someone was hit with that spell, while the 95% is the percentage amount they use the spell (number of that spell's hits/total of all spell hits). For shamans, I'd expect to see a lot of Chain Heals, just like we see here.
Average amount of healing per tick is a good reflection of gear, buffs, and on use +healing trinkets. This is because those directly influence healing per cast output. If you're not sure how undergeared a new healer is, this is the place to look. What you should see over time is new guildies getting closer and closer to your veteran players. It can also explain seemingly random gear discrepancies between players. For example, if a certain healadin forgets to take off their resist gear after Mother, you're going to see some sad numbers between there and Council. Not that it's happened... /shady eyes Little differences add up over a raid night.
Crit only matters to Shaman and Paladins. A while time ago I covered good bench marks for crit. You can find them here.
The number of hits or ticks, depending on the spell, are a good litmus test for how engaged a healer is. In a lot of T5 content and most of T6, you need proactive healers who are always looking for the next bit of damage. You don't want to see someone that's waiting around for damage, or worse, trying to save mana by having others heal their assignments. In our example, both Shamans are very close to each other with respects to number of hits, although Hwesta does come out slightly ahead. What you don't want to see here is something like half the hits compared to another of the same class (remember, we're talking about people that were present in the raid for the same amount of time.) In that past, something like that indicated low trash healing participation.
On Wednesday we're going to cover how to break down an individual fight and how to tell who's healing on trash and who's watching tv. Busted! Hope to see you then!
Marcie Knox continues to be the healing lead for her guild and wishes more people hated Summer. That way they'd show up for raids instead of going to mysterious things called "BBQ's". Don't forget she's always looking for nifty healing screenies (marcie[dot]knox[at]weblogsinc[dot]com).
Filed under: Druid, Paladin, Priest, Shaman, Analysis / Opinion, Tips, How-tos, Raiding, Raid Rx (Raid Healing)






Reader Comments (Page 1 of 1)
Faerun Jun 23rd 2008 5:56PM
Great column! Is there any similar one out there in the blogosphere that deals with WWS parsing for DPS and Tanks as well?
Thanks!
Daddywarbuck Jun 23rd 2008 7:01PM
I disagree completely.
How much you heal is completely dependent on who you're supposed to be healing, who actually looks at healbot to make sure the person needs healing and who just clicks healbot faster.
It also depends largely on how much damage the raid is taking (duh).
The *only* time these numbers matter is when you're not getting enough healing and dmg over time is greater than cast of a heal+1 second (reaction time + lag).
The rest of these times, heal percentages are often bunk. The only critera for healing is: Did the people who took damage from playing reasonably well get healed as much as they should of in the situation.
--Daddywarbuck of Farstriders
Shumina Jun 23rd 2008 8:15PM
I gotta disagree with you. While that strategy may work for your raid's healing, it'd be an invitation to /rollface-town in mine. WWS backs up the hunches when you think you're having to shoulder the weight of healing more than you think is reasonable because someone's coattailling.
Someone sucking? We mention that all healers need to wake up. Then after the raid, we look at the numbers (taking assignments into consideration) and always always always the hunch of who's slackin' is proved correct.
Thanks for these posts and keep 'em up!
Daddywarbuck Jun 23rd 2008 9:23PM
As I said:
The *only* time these numbers matter is when you're not getting enough healing and dmg over time is greater than cast of a heal+1 second (reaction time + lag).
So there it matters
The rest of the time, heal numbers are BUNK.
Often DPSers look at them and make judgements when really the numbers are unimportant. For instance, DPS and damge numbers are always important, even when you're doing a lot of CC for your class.
So many DPSers look at WWS and say WOW _____ healed 44% of the raid, blah blah blah is sucking cause they only did 17%. Really what happened is that blah blah blah didn't case extra heal spells because they weren't needed. and did her job just fine.
WWS is *ONLY* for finding problems. Not for rating raids where the healing was satisfactory.
Extremitus Jun 24th 2008 3:29AM
Shumina, I'm of the opinion that if you take the correct numbers of healers to a raid, that is to the say the bare minimum, then the healers won't notice if they're shouldering more weight - if someone doesn't pull their weight then someoene dies, and you can usually tell from the assignments and grim reaper instantly who it is that isn't doing their job. Not that WWS doesn't do a hell of a lot for post-raid analysis, I'm just funny about healing meters I guess.
Rihlsul Jun 24th 2008 8:32AM
Exactly - this is in theory a way to find sub-par performers. If you're just starting out on an instance, it's common to travel a little healer heavy. As the raid learns the encounters, the healers should notice they're having to work less, and you should be able to swap in a DPS or two - making the whole run go faster and smoother.
Being able to spot the sub-par healers to swap out is a great way to manage that.
Daddywarbuck of Farstriders Jun 24th 2008 2:11PM
@ Rihlsul
% does *not* spot the sub-par healers when you're not at healing capacity.
Discretion is often the better part of valor. Frankly if I find out a driud is *EVER* coming close to me in single target healer, there is a big problem with that driud. Healing less is often the correct answer.
This would be different if the only cost to heal is your time, but for shamans and paladins, mana is a real cost, and prudence in choosing when to heal is a big decision.
Your position that the HPS meter or total health healed meter is useful in figuring who to not bring when you start cutting healers is in error. The guy not healing is often keeping a smart reserve which can be used in the case of mistakes or other exigencies to use time efficient but mana inefficient heals.
The healers job is *not* to max their heal per second. It is to keep people alive. The best way to do that is often contingency planning and not running down to the wire.
Silverrealm Jun 26th 2008 11:15AM
Defo agree with Daddywarbuck on % not = sub-par healers.
I am druid and I use Heal Over Time heals. Depending on group make up, experience, gear, number of healers and a few other factors... my % might be low for some encounters and through the roof on others.
% is not the end-all be-all of telling who is a good healer and who isn't.
If I am on group heals and the MT and OT have their own healers... sometimes my ticks are insanely high, depending on the fight... other times the tanks take ALL the damage and I rarely have to heal the DPS or healers, so my numbers would be very low.
Wyred Jun 24th 2008 8:25AM
Sorry daddywarbucks, but gotta disagree with you on 2 points. The first is just a technical one. As a holy priest I always have a greater heal casting and stop if it's not needed, a very common practice for my class, so cast-time + reaction time doesn't neccessarily apply. The second is, while healing may have been satisfactory for a raid, if someone *is* doing less than they could (and note the OP is only comparing same class/assignments/gear), while it may have been good enough for that raid, it could be what's holding you back in new content. WWS can help you look at spell rotations, heal per tick and so on to find out what it is they might be doing wrong. Healing meters are useful, they're just more difficult to interpret than dps meters. Which is the entire point of this column
Smon Jun 24th 2008 11:31AM
Marcie, thank you for your awesome column, I have been following it for quite some time, and it has definitely helped me improve my game as well as analyze the other healers I play with.
I think meters are extremely useful in post fight analysis. While I agree that it is probably more prudent not to compare from one class to the other, I have found that what I can compare is each players' position relative to the others on the meter for a given fight. Taking into account what kind of healing the fight requires I have a decent idea of where to expect the healers I play with to dial in. When I am pleasantly surprised by someone's big jump in rank, I take a closer look at what they did for that fight. I generally also use the meters after every fight to:
1. Assess whether I am healing at the level I feel I should be at.
2. Make sure the rest of the healers are each carrying their weight.
I also wanted to point out that I have found that play style is really the thing that has the greatest effect on healing output. There are people who while being undergeared can consistently and easily outheal some of the other healers simply because they are better players.
Lastly @ Daddywarbuck if one healer puts out 44% of raid healing and the other only does 17%. Even if no one died, there is something very wrong with this picture and the second player is in error. Healers should never be healing just enough, they need to be proactively casting and react quickly to the situation. Also, no healer should rely on the other healers to put out much more healing than anyone else, this leads to problematic situations where the big healer gets stretched too thin. In a situation with such a large healing discrepancy, the 44% healer is probably stretched to the limit, which is not very prudent if a large amount of burst dmg goes out. (as the other healers expect this person to do most of the healing) I always play like no one else is casting on my target, and if I see them get healed before my cast goes off, I cancel my spell and move on to the next player. I expect my fellow healers to play the same way. It establishes a kind of redundency which leads to fewer lost lives and more dead bosses.
Daddywarbuck of Farstriders Jun 24th 2008 2:10PM
In response to:
++
Lastly @ Daddywarbuck if one healer puts out 44% of raid healing and the other only does 17%. Even if no one died, there is something very wrong with this picture and the second player is in error.
++
Not at all. Raid boss encounters have all sorts of behaviors for which a crazy large unused pool of mana which can be used for the maximal time-efficient heals. Keeping that large pool of mana in the case when one of those bad, burst damage things happens, is important. Its just like keeping buffer money in savings accounts instead of putting it all in a savings account. Sh*t happens, and you better have a shovel when it does.
++
Healers should never be healing just enough, they need to be proactively casting and react quickly to the situation.
+++
"just enough" is exactly how much healers should be healing. That's why its called just enough. Just enough is usually "out of instant smooshing" range for squishy low stamina casters, and up to about 90% for tanks and other people who "should" be getting hit. I think you'll find you're doing that anyhow, as that is what most heal plugins are geared to warn you about.
If there is free mana that can be used to heal people up more, then it should be used in the most mana efficent way you have time for to get them up near full health. Unless there is a big hit possibility, or you're going to max on mana, no one should ever be healed above 95% of their health unless you have tons of mana to spare.
+++
Also, no healer should rely on the other healers to put out much more healing than anyone else, this leads to problematic situations where the big healer gets stretched too thin.
+++
Actually, that's efficient use of +heal. The big healer should be pumping out as much as they can without overhealing should let them. The less big healers should be conserving mana in case the big healer fails, or dies, and should be topping off. This is the way to get the most time economical healing possible without running out of mana in all healers. Additionally, when the big healer DOES run out of mana, the healers can reverse and then the main healer can be the sippy healer, and the guy who was just topping off can be the main. This happens in my raids when the guy throwing mana around like crazy goes /oom.
+++
In a situation with such a large healing discrepancy, the 44% healer is probably stretched to the limit, which is not very prudent if a large amount of burst dmg goes out.
+++
It is EXTREMELY optimized for that situation, and very prudent. Your raid will have ample mana to heal everyone with the scariest large heals ever. I'm not talking about the bare metal, learning the fight, when all the healers are gritting their teeth and almost constantly casting, I'm talking about the middle ground before you start cutting healers and after you get out of the die 20 times a night period.
+heal adds to your heals. If you go over the person's max hit points, that mana is lost. If the raid takes 6k damage each, the fact I can cast 10 rank 11 holy lights and heal everyone back up is going to be a turning point (if I'm the guy who's not on balls to the walls healing duty).
+++
(as the other healers expect this person to do most of the healing) ..... I expect my fellow healers to play the same way. It establishes a kind of redundency which leads to fewer lost lives and more dead bosses
+++
And this leads to the big issue you're facing: You're expecting a strategy for everyone to do the exact same thing. That is an unsophisticated strategy, which will work via enough gearing, does not make use of conscious resource management. It ends up with everyone staring at bars and no one really looking out for where critical aid is needed, or rather everyone running that way and then neglecting everyone else.
It is *much* more mana and hps economical for there to be people who are running at full speed, and several other who are dealing with crises only, but otherwise observing the 5 second rule. You SHOULD see a vast imbalance in the healing on the nights when emergencies don't happen and that extra capacity didn't happen to be needed to fix some screwup or another.
Just like everyone should not be 100% busy in a company, not every healer should be 100% casting. Just like in a company, too many people not healing should mean layoffs, but some people not working some of the time is the cost of having a reactive business (or raid), and is a good thing!
Isambaard Jun 24th 2008 12:45PM
Small nit to pick. Crit is utterly unimportant for Shaman. The buff is neat and all, but since we're virtually never tank healing and the buff is virtually useless to everyone else crit is pretty useless also.
Some of our gear has it, some doesn't. Should we take the paladin healing shield off Supremus instead of the shaman healing shield out of Hyjal? Of course not, Paladins actively benefit from crit, its a key component of how they last in long fights. For shaman mp5 is the key and so we should always prefer mp5 gear over crit gear.
ScottyO Jun 24th 2008 5:30PM
@Daddywarbuck of Farstriders :
So, let me test my understanding. Lets say your raid has 8 healers. You would rather end the fight with 4 healers at 0% mana and 4 healers @100% mana than to have all 8 @50% mana? Thats probably one of the stupidest things I have ever heard (when it comes to raid healing). If the work is divided evenly, and all of a sudden an emergency happens, wouldn't 8 people flash healing (then digging into their mana regen bag of tricks) be better than 4 people doing it? As a healing lead in a black temple guild, I say yes.
ScottyO Jun 24th 2008 5:39PM
@Daddywarbuck of Farstriders : Also, your comments about healing do not seem to really fit into the context of this post. Marcie is helping us break down healing strats using tier 6 content as an example. Believe me, you have not healed until you have healed tier 6. Healing is a whole different game from your guild (karazhan cleared and a couple bosses in ZA). Not a single 25 man kill yet, and this is clearly targeted at more progressed healing. Your posts are lengthier than the article itself, and frankly I would rather take healing advice from the healing lead of a t6 clearing guild than yourself who has cleared nothing but Karazhan.
Flooq Jun 25th 2008 12:38AM
Interesting that you've looked at how much content Daddywarbuck has seen. I could tell from the comment about tanks sitting on 90-95% that it wasn't much.
This isn't an attack on him, his views seem perfectly reasonable for someone who has only seen some 10-mans. However, reactive healing will not work much further into the raid content.
1) In lter content you do not want your tank sitting on 90%, you want heals in the pipeline constantly so that they're already halfway cast by the time he takes a damage spike.
2) No T6 fight is going to let your healers take turns at doing the work. If they can't work together at the same time without going oom then they must have extremely low regen stats, perhaps they are undergeared or perhaps they're just too stacked on +heal.
3) No fight that throws 6k damage at the raid in a short space of time is going to give anyone a chance to cast 10 holy lights.
4) Bonus healing isn't for letting you take more of the burden, it's to help you heal on the same scale as damage the instance is throwing at your raid members.
5) Yes, there is a point when people can heal too much or simply use inefficient but quick spells to make themselves look good on the meters. Again an indepth analysis like this article describes will allow an experienced raid/healing lead to spot that problem.
Teresa Jun 25th 2008 3:58PM
You can't JUST look at WWS.
For example, we run 2 CoH priests, and 1 IDS priest. IDS priest is assigned to tank healing, while the 2 CoH priests are assigned to melee/raid.
The IDS priest will have low numbers (5-10% effective healing, with 50%+ overhealing due to spamming a downranked gheal), while the 2 CoH priests will both have 20%+ healing each. By your 10% factor, the IDS priest wouldn't be "Pulling their weight" and would come under close scrutiny, even though they were doing their job, and their assignment didn't die.
Honestly, a dps should NEVER be looking at the healing meters and criticizing healers. If theres an overall healing lead, THEY should be the only ones looking at numbers, and seeing if they're rational or not. If theres not, the priest/shammy/druid/pally leads should be looking at the numbers, seeing if they're reasonable considering their assignment, checking to see how many times they came off their target to heal someone else (other than themself), and whether or not their healing assignment died.
Unless tank targets are dying, ppl really shouldn't be criticizing or second-guessing the healers.
Wyred Jun 26th 2008 10:55AM
@15, you seem to have misinterpreted the original post. It's *not* comparing the tank healing imp DS priest to the CoH healers. If you look at the comparison made in the article, the obvious analogy here would be comparing the 2 CoH raid healers. If their gear is similar but there was a big disparity in healing, there might be something else going on. WWS would then help you look at what that might be. Obviously it's not the only measure of whether healing was effective or not, but it is useful when interpreted. I'm not sure where the dps comment came from. As for for 'if they live, then the healing was good enough' not neccessarily. What if you're downing the boss 33% of the time because you get a bit luckier with crits, crshing blows etc? Is the healing good enough then? To my thinking in a situation like that analysis, using a tool like WWS, can help you put that boss on farm.