1. #1

    r shaman numbers and reading raidbots

    Just perusing raidbots and looking at resto shaman pure numbers, we are bottom of the barrel across the board in 10 man heroics and just slightly ahead of only resto druids in 25. I was mostly looking at all parses as this seems more indicative of the general healing capability. It seems like so many of the top 100 parses involve a healer being dead for half the fight and thus give strangely inflated numbers. What do ya'll look at? Do these numbers really mean much at all? With resto druids receiving a probable 10% buff, are we now at the low end of the totem poll?

  2. #2
    RShaman are shit on farm.
    Signature size violated forum rules.

  3. #3
    Quote Originally Posted by Ushra View Post
    RShaman are shit on farm.
    Ok great thanks for that. I realize that mastery is devalued in farm content as healers are taking care of damage quicker but raidbots is data. If everyone is at max potential (farm), shamans are at the bottom. Ultimately I realize that we are more useful when people are fucking up but by the time we down most heroic fights everyone is pretty much on top of their game.

  4. #4
    Moderator Endus's Avatar
    Join Date
    Feb 2010
    Location
    Waterloo, ON
    Posts
    20,944
    DPS meters really are a good measure of DPS performance; their contribution is only measured in how much health they take off the boss, in a given time frame.

    HPS meters aren't a good measure of healer performance for a host of reasons;


    • "Effective healing" means healing the target that needs it, in the time frame they need it within. If a guy's at 90% health and isn't going to take more damage, and has a HoT from another healer already on them, tossing your own HoT on them is a waste of mana and provides zero effective healing, though it'll show up on HPS meters, padding your numbers while lowering the other guy's.
    • As mentioned; if one healer dies, it gives another healer a lot more health to heal, increasing their apparent effectiveness. This is like using AoE numbers for DPS valuation, though.
    • As gear improves, DPS increases, but healing often decreases, due to players getting better at avoiding shit, and because everyone has more HP and healers have higher potential HPS, topping people off faster. In an ideal world, adding more DPS makes the fight shorter and everything easier. Adding more HPS does absolutely nothing unless people were dying from lack of healing.
    • Resto Shaman are particularly vulnerable to the latter, since their Mastery is tied to play health, and both increased health pools and higher overall HPS from all healers work together to make Resto appear artificially low. The issue with this is they still have the potential to do much better, but because the farm content is easy, nobody's in enough danger to get them to push. So not only do healer mechanics in general punish healer performance over time, Resto Shaman mechanics also punish Shaman players additionally.


    The important thing to get out of all that is that HPS meters are a really, really bad way to measure healer effectiveness. They aren't nearly as reliable as DPS meters are.

  5. #5
    Thanks so much for the response. Since HPS meters are a bad way to measure effectiveness, what would be a good way?

  6. #6
    Moderator Endus's Avatar
    Join Date
    Feb 2010
    Location
    Waterloo, ON
    Posts
    20,944
    Quote Originally Posted by rmlunsford View Post
    Thanks so much for the response. Since HPS meters are a bad way to measure effectiveness, what would be a good way?
    The only really accurate way is to study effective healing. The difficulty is that it's impossible to stick that on a simple graph, like you do with HPS or DPS numbers, since it relies on both the gap between damage received and healing applied, and the fight-specific needs for healing according to which good healers are apportioning their casts.

    If you're looking for a site that shows you a graph, it's not possible. It needs to be done on a per-log basis, by hand, with a solid grasp of the fight in question and how every healer's mechanics work.

  7. #7
    Is Endus' response something everyone agrees with? I can't see how healing meters are useless. You have a certain amount of damage that a raid takes during a fight and it needs to get healed. Yes it varies but that damage still needs to get healed and if you heal more of it you're doing a better job. I understand your points but I personally think that if you're not trying to win you're not really pushing yourself. Obviously, you don't want to cast healing rain because you'll get more hps out of it when the tank has 20k health and he needs a heal. At the same time, If we are on the same fight and I do 30k and you do 50k, I'd wager that 95% of the time when you go back and do your log analysis you'd see that you did a much better job than I did.

  8. #8
    The Patient edlike's Avatar
    Join Date
    Jul 2011
    Location
    Orlando, FL
    Posts
    260
    Quote Originally Posted by rmlunsford View Post
    Is Endus' response something everyone agrees with?
    As far as raidbots is concerned, absolutely.

    No one is saying healing meters are useless. It is an effective tool to quickly analyze what's going on in the raid. Concerning rankings and raidbots though, healing meters are very misleading compared to DPS meters for all the reasons endus mentioned.

  9. #9
    Moderator Endus's Avatar
    Join Date
    Feb 2010
    Location
    Waterloo, ON
    Posts
    20,944
    Quote Originally Posted by rmlunsford View Post
    Is Endus' response something everyone agrees with? I can't see how healing meters are useless.
    They aren't useless. They're just not an accurate measure in and of themselves. It's like looking at dummy DPS meters, or AoE trash spam DPS meters. There's data there, and it DOES have meaning, but if you're trying to use that as a measure of DPS spec rankings, it's going to be garbage.

    You have a certain amount of damage that a raid takes during a fight and it needs to get healed. Yes it varies but that damage still needs to get healed and if you heal more of it you're doing a better job. I understand your points but I personally think that if you're not trying to win you're not really pushing yourself. Obviously, you don't want to cast healing rain because you'll get more hps out of it when the tank has 20k health and he needs a heal. At the same time, If we are on the same fight and I do 30k and you do 50k, I'd wager that 95% of the time when you go back and do your log analysis you'd see that you did a much better job than I did.
    That's exactly the issue.

    There's two variables to both damage-dealing and healing. One is the straight oomph; the healing you can push, or the damage you can deal. The other is the time factor.

    DPS is simple; there's a single HP pool that you need to reduce to zero. That's your sole goal (barring a few exceptions). Because that's your only goal, more DPS means that HP pool hits zero faster, meaning you change the length of the fight.

    Healing is more complex. Everyone has two critical points; 100% health, and 0% health. At 100%, healing done is wasted. At 0%, they die. Your goal, as a healer, is to maintain everyone at some level between 0 and 100. Damage taken per second in a raid fight stays pretty much the same (again, barring some exceptions, like HP percentage triggered AoEs). The first goal in healing is to ensure nobody hits 0%. Once you can manage that, nobody dies. Then, for safety, you try and maximize people's time at 100%, to provide greater wiggle room. This is much less valuable, in the long run, than DPS making the fight shorter; that reduces the damage people will take, and has a greater effect on survivability in the long term, once healers can keep everyone up for the duration. Improving on how close to 100% you can keep people is nice, but it doesn't bring anything of direct value to the raid; if nobody's dying anyway, you're not making the boss kill any easier, save that other players can get lazier about avoiding damage. With Resto Shaman, this is additionally complicated by our Mastery, which is more effective the more difficult the fight is, and the closer to zero health more people are.

    Basically, Resto Shaman end up doing less healing as people gear up not because Resto Shaman aren't capable of pushing as hard, but because the content is getting so much easier they don't get the opportunity to push any harder. It's automatic, which means they can't artificially bump their performance by popping CDs for shits and giggles, like other healers can. Because, again, once everyone's surviving, the extra healing is just "nice", not directly beneficial in any concrete way.


    HPS is a part of healer effectiveness. But it's not the whole story. DPS really is the whole story for damage-dealers, though. That's why you can't just use HPS meters the same way you use DPS meters; because it's only part of what you need. And there's no easy way to turn the other aspects of healing into raw data to make it easily comparable like that.

  10. #10
    Marvelous. Thanks so much for taking the time respond so thoroughly. It's much appreciated. Since you've so eloquently put to rest the value of hps as a viable metric, how does anyone go about determining whether a class needs an increase? For instance, Blizzard is buffing Druids. What sort of metric do they use to determine that the buff is needed?

  11. #11
    Moderator Endus's Avatar
    Join Date
    Feb 2010
    Location
    Waterloo, ON
    Posts
    20,944
    Quote Originally Posted by rmlunsford View Post
    Marvelous. Thanks so much for taking the time respond so thoroughly. It's much appreciated. Since you've so eloquently put to rest the value of hps as a viable metric, how does anyone go about determining whether a class needs an increase? For instance, Blizzard is buffing Druids. What sort of metric do they use to determine that the buff is needed?
    I have no idea how Blizzard does it.

    As far as players determining this, a lot of it comes down to "feel". If you're pushing new content, and you're just not able to keep up in healing, and swapping to another class/spec is significantly better off in most fights, then there's an issue. Once you repeat that kind of testing across enough people who know what they're doing to ensure it's a class issue, not a player issue, you've got something to move forward with.

    PvP discussion works much the same way, since specs need to be evaluated on the total package they bring to a team.

  12. #12
    On the topic of "use of raidbot for healer", couldn't the "sample" mesure (especially on hard modes) be slightly usefull ? I know popularity is not the same thing as "usefullness", but can't we still assume that if a spec seem "strictly" worst than another, they'll appear less often on kills ? Or do you think that this argument is totally useless even when restricting ourselve to "hm kills" ?

  13. #13
    Dreadlord zshikara's Avatar
    Join Date
    Jan 2011
    Location
    Arlington, TX
    Posts
    993
    Quote Originally Posted by rmlunsford View Post
    Is Endus' response something everyone agrees with? I can't see how healing meters are useless. You have a certain amount of damage that a raid takes during a fight and it needs to get healed. Yes it varies but that damage still needs to get healed and if you heal more of it you're doing a better job. I understand your points but I personally think that if you're not trying to win you're not really pushing yourself. Obviously, you don't want to cast healing rain because you'll get more hps out of it when the tank has 20k health and he needs a heal. At the same time, If we are on the same fight and I do 30k and you do 50k, I'd wager that 95% of the time when you go back and do your log analysis you'd see that you did a much better job than I did.
    healing isn't a game that can be "won" unless everyone in the raid is still alive at the end of an encounter. I'm oversimplifying it, because the entire game is healing to me. I may not be as vocal about it as endus is, but I will never play anything other than a resto shaman. From back in BC when I was a chain heal bot, to wrath when I was actually hard haste capped (LHW had a natural 0.98 second cast time), to cata when mastery came into the game making us what we are today, I will always be shaman. Endus is 100% right in every way. We are truely a utility healer, not a raw numbers healer like a paladin or priest.

  14. #14
    Quote Originally Posted by Bethan View Post
    On the topic of "use of raidbot for healer", couldn't the "sample" mesure (especially on hard modes) be slightly usefull ? I know popularity is not the same thing as "usefullness", but can't we still assume that if a spec seem "strictly" worst than another, they'll appear less often on kills ? Or do you think that this argument is totally useless even when restricting ourselve to "hm kills" ?
    That doesn't really work as well because stacking healers is rarely required. Healing damage is still healing damage, and most raid groups will not switch out a healer to a different spec unless there are specific boss mechanics to account for.

    If raidbot numbers would have been the be all end all of performance, you would never see resto shaman+ resto druid healing a fight, or double disc being the only two healers in the raid.

    Also, I would personally say that player skill plays a greater role when talking about healers, rather than dps. Whereas a dps rotation is pretty much always constant and can be learned on a reflex level, no healing encounter is the same and your snap decision making skill plays a greater role.
    Last edited by Blindlad; 2013-01-22 at 04:18 PM.
    Healing with different classes since Vanilla
    http://eu.battle.net/wow/en/characte...Higgs/advanced

  15. #15
    Quote Originally Posted by Blindlad View Post
    That doesn't really work as well because stacking healers is rarely required. Healing damage is still healing damage, and most raid groups will not switch out a healer to a different spec unless there are specific boss mechanics to account for.

    If raidbot numbers would have been the be all end all of performance, you would never see resto shaman+ resto druid healing a fight, or double disc being the only two healers in the raid.

    Also, I would personally say that player skill plays a greater role when talking about healers, rather than dps. Whereas a dps rotation is pretty much always constant and can be learned on a reflex level, no healing encounter is the same and your snap decision making skill plays a greater role.
    Raidbot numbers aren't the end all of performance but they are 90% of what hardmode guilds use to guage a players performance. I was recruited to a top 100 guild because they saw my logs on WOL.

    Furthermore when i joined them my healing took a major dip because they roll double disc and hpally.

    what your seeing is what long time players call mid season shaman performance. Healers and DPS are geared enough so that your not surviving by the skin of your teeth anymore and our mastery goes to the toilet. Until our mastery changes to something more stable you will see this in every expansion. If you want to top the healing meters wait until the final tier where damage and healing skyorckets and the encounters become big stack up and see how you can do again. Usually the second raiding tier is a bunch of movement and weird fights so why not try leveling a priest or a pally?

  16. #16
    Quote Originally Posted by Endus View Post
    DPS meters really are a good measure of DPS performance; their contribution is only measured in how much health they take off the boss, in a given time frame.

    HPS meters aren't a good measure of healer performance for a host of reasons;


    • "Effective healing" means healing the target that needs it, in the time frame they need it within. If a guy's at 90% health and isn't going to take more damage, and has a HoT from another healer already on them, tossing your own HoT on them is a waste of mana and provides zero effective healing, though it'll show up on HPS meters, padding your numbers while lowering the other guy's.
    • As mentioned; if one healer dies, it gives another healer a lot more health to heal, increasing their apparent effectiveness. This is like using AoE numbers for DPS valuation, though.
    • As gear improves, DPS increases, but healing often decreases, due to players getting better at avoiding shit, and because everyone has more HP and healers have higher potential HPS, topping people off faster. In an ideal world, adding more DPS makes the fight shorter and everything easier. Adding more HPS does absolutely nothing unless people were dying from lack of healing.
    • Resto Shaman are particularly vulnerable to the latter, since their Mastery is tied to play health, and both increased health pools and higher overall HPS from all healers work together to make Resto appear artificially low. The issue with this is they still have the potential to do much better, but because the farm content is easy, nobody's in enough danger to get them to push. So not only do healer mechanics in general punish healer performance over time, Resto Shaman mechanics also punish Shaman players additionally.


    The important thing to get out of all that is that HPS meters are a really, really bad way to measure healer effectiveness. They aren't nearly as reliable as DPS meters are.
    PreCata I'd agree with that, but game has becomes such an AoE dmg based design. Theres a lot of outgoing dmg going on that can't be avoided while this used to be less the case.

  17. #17
    I'd say that pure hps throughput is a lot more important than some of you are making it. I'm sure we've all experienced the feeling when you're working on a difficult progression encounter, and a mediocre healer (throughput wise) gets switched out for a really good one, and suddenly healing becomes 10 times easier.

  18. #18
    as a 10 men progress semi hardcore, healing shaman i cannot remember having someone die to a bosses ae dmg due to too little raid hps in a VERY long time (tbc?), as the healers and dd i play with now what to do and how.

    The most death nowadays come from critical mistakes (omnotron defense system style, bot council style, heroic will spark soaking "ups i forget to pull my cd", getting owned by windlords bladefurry, taking a green circle from bladelord and things like this), if something is healable / meant to be healable from the encounter mechanic side, no one dies.

    Only example i had in current content of "this guy died in between 2 globals of me spamming my highest hps heal" was sometimes tank death at pretty high stacks at heroic lei shi.

    In my group, the hps itself is lowest in the most encounters, running with a discipline priest and a paladin or a monk. It doesnt validate the power of the cd usage, the power of the 10% hp buff i give people, and the power of the healing i could provide if things wouldnt go so well.
    But i cant attack this 300 k absorbshield every minute on the whole group (well exept telling people to macro every spell with a (/cancelaura spiritshell).
    I cant compete with ridicoulus attonement healing on "this boss now takes 600 % more dmg for a quarter of his life expectancy"
    I cant compete with monks renewing mist + uplift spam on hps term. but i dont need to.

    I as a shaman can provide enough hps to cause no one dying to events that are meant to be healable.
    The only reason for demanding a shaman buff is to be able to compete with op monk / disc.

    Paladin / shaman / druid are okish (perhaps druid a little on the weak side, but same as shaman druids hps numbers are suffering from absorb crazyness discs + uplift spam), while the other 2 specs are over the top atm. disc will be inline with 5.2, and monk will perhaps one day too get in line, but its difficult to bring monks in line, as they have very little raid toolkit, exept hps spam, which is all monk is about.

    about the usefulness of hps to measure "skill" of a healer:
    I remember my first stoneguard hc kill, in ilvl 470, somewhen in december or november.
    i pulled 80 k + hps, everyone was near dieing all the time. Tanks eating dmg like mad. Boss going in enrage, iceblock + dots kill it. My healingrain overheal is 3 %.

    Now, i have ilvl 500+ and i can maximum pull 50 to 60 k hps in this fight, and everyone is over 80 % nearly all the time. I have 75 % overheal on healingrain. Possibly i could even heal it alone. , but i have no dire need to do this.
    Last edited by Holofernes; 2013-01-23 at 04:51 AM.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •