Just a bit offtopic, im attempting first bosses of SSC, TK an Magtheridon. Im stacking +healing and mp5 and a bit of crit so i can get at least to 19% crit rating unbuffed. With that in mind i get 1900-2000 healing and 114 mp5 unbuffed. But now i see the top paladins in my server in ther armory with yeah, lots of healing (2400+) and crit (27-30%) BUT less mp5 than i thought (70-80 mp5). How come they need so few? is that because the use flask/consumables of mp5? I really dont get it. Can you explain me when do you change from mp5 stacking to crit stacking?
P.D.: Yeah, i have read the whole thread this past months and still feel curious why top class paladins stack crit so much.
Just a bit offtopic, im attempting first bosses of SSC, TK an Magtheridon. Im stacking +healing and mp5 and a bit of crit so i can get at least to 19% crit rating unbuffed. With that in mind i get 1900-2000 healing and 114 mp5 unbuffed. But now i see the top paladins in my server in ther armory with yeah, lots of healing (2400+) and crit (27-30%) BUT less mp5 than i thought (70-80 mp5). How come they need so few? is that because the use flask/consumables of mp5? I really dont get it. Can you explain me when do you change from mp5 stacking to crit stacking?
P.D.: Yeah, i have read the whole thread this past months and still feel curious why top class paladins stack crit so much.
Because a shadowpriest is what, 250 mp5? Chain potting is another 100, taking raid damage and getting healed (a big factor in all sunwell fights) is easily another 30-35 mp5.
I mainly changed when I started potting more and had enough +heal to make FoL a very strong spell. This was t6 content for me, but I kind of skipped half of the t5 content.
Paladins with 2400+ heal dont really stack crit btw, that kind of number is only achievable with using 22 healing gems. 27% crit is pretty easy to get without gemming any crit.
I don't think it is valid to ignore it then either. If you are doing your maximum HPS throughput you are spamming HL11. In that situation most of the time someone dies it is because your heal didn't get there in time. Say it did land in time but the target still died, it is not likely it would be different it it healed for 100 more. You would have to be lucky and have the numbers work out well for the extra 100 healed to keep the target alive. While in a lot more circumstances if your heal landed and the target still died, he wouldn't have if the heal did crit. So I would say the increased chance of the heal critting and saving the target is just as (if not more) likely then the target not dying because your minimum healed was for 100 more.
Unless you're going from 99% to 100% crit, adding crit does nothing for your worst-case burst, which is what matters for avoiding tank death. In general, you shouldn't be trying to reduce the chance that the tank dies due to insufficient healing, you should be trying to eliminate it. If that's impossible then crit's burst value is non-zero (and is probably higher than +healing), but if it's impossible it nearly always means that your raid is doing something wrong (Brutallus with an avoidance tank would be the most recent exception).
Unless you're going from 99% to 100% crit, adding crit does nothing for your worst-case burst, which is what matters for avoiding tank death. In general, you shouldn't be trying to reduce the chance that the tank dies due to insufficient healing, you should be trying to eliminate it. If that's impossible then crit's burst value is non-zero (and is probably higher than +healing), but if it's impossible it nearly always means that your raid is doing something wrong (Brutallus with an avoidance tank would be the most recent exception).
I think you missed my point. Most of the time when your heal hits the tank and he still dies, he wouldn't have survived if the worst case heal was 100 higher (for example). But if you heal hits the tank and he still dies, he is very likely to survive if the heal would have crit. Enough so to say 5% increased chance to crit is just as likely (maybe more) to save the tank in that situation.
I don't think saying that increasing worst case heal could eliminate it, while crit will only reduce it is really valid. Because in any realistic scenario of choosing between values +heal or crit, the change of +heal won't be enough to eliminate a worse case tank death.
I am not saying that +heal is useless, it is the stat I mostly stack. Just that crit shouldn't be ignored while considering burst HPS.
Just a bit offtopic, im attempting first bosses of SSC, TK an Magtheridon. Im stacking +healing and mp5 and a bit of crit so i can get at least to 19% crit rating unbuffed. With that in mind i get 1900-2000 healing and 114 mp5 unbuffed. But now i see the top paladins in my server in ther armory with yeah, lots of healing (2400+) and crit (27-30%) BUT less mp5 than i thought (70-80 mp5). How come they need so few? is that because the use flask/consumables of mp5? I really dont get it. Can you explain me when do you change from mp5 stacking to crit stacking?
P.D.: Yeah, i have read the whole thread this past months and still feel curious why top class paladins stack crit so much.
It depends. I still have around 150mp5 unbuffed set for council fight without SP/shammy - lately we have 0 resto shamans and only 1 SP in caster DPS group - and for gringing purposes. If you assume you can get SP in your group than you can lower your mp5 to any amount possible for reasonable pally gear (say 70-80, but dunno if you can have 70 in pally (aka no spirit) t6 content gear). My calculations showed me that with alchemist trink you shouldn't worry about mp5 at all (in average t6 fight with mp5 consumables and chain chugging mana pots) even without SP. If you don't have alch stone then "optimal" number (calculated from averaged mana needs for usual t6 fight) is around 120-130 mp5 unbuffed. That can be easily obtained with one good mp5 trink and usual +heal socketed gear.
Crit's not valuable for surviving worst case scenarios. But that doesn't mean that crit has no value.
No mob does completely predictable damage. Normally you have a range of damage.
Say that a boss has 2000 outgoing DPS on average. But assume a good avoidance streak would lower that to 1000, and a poor avoidance streak would raise that to 3000.
Now, when you're assigning healers, you wouldn't want to assign enough healers to pump out a constant 3000 hps, because unless you overgear the content, that would be a poor allocation of resources. (Imagine the healing needed if you automatically assumed that brutallus landed every single swing).
You're working with scarcity, and so you tailor your healers to be able to recover the average amount of incoming DPS.
So what happens when that tank takes a period of low avoidance? His health lowers and hangs lower than full, because he's taken more damage for a short time that the healers weren't able to keep up with. I'm sure you have noticed this phenomenon. You've got healers healing the tank, but his health never quite tops off.
This is where crit heals are helpful. While random streaks of avoidance will temporarily lower the tanks equilibrium point, crit heals will bring it back up. Now you may argue that you can get other healers to help top off the tanks in those situations, or use cooldowns like NS to catch them back up, but again, unless the fight is trivial, there is a scarcity, and why would you divert resources, or cooldowns to do something that can happen passively through spell crit?
The difficult thing is, when spell crit is doing it's job, you don't even notice it. The tank gets lower, but moves back up to full health because a holy light crit. Were that holy light to not crit, the tank probably wouldn't die outright, but he would hang a few thousand health lower than his max for a while until someone else took some corrective action.
You will rarely see it saving the day. It's good for stability, and it offers a form of increased efficiency. It's an important stat. But I think mostly it's over-calculated, either distilled into purely a regen stat, or occasionally an HPS boosting stat. But personally, I think of it as a good all-round stat, increasing tank stability, efficiency, and occasionally raw HPS.
To try and really distill it is kind of cheating it. It's nebulous, because it's based off the RNG, which in general is not reliable for healing. But it is useful enough to help counter the opposing target's RNG abilities, and reduce the frequency where you have to use alternative cooldowns and resources, which are limited, to battle "bad luck streaks".
I don't think it is valid to ignore it then either. If you are doing your maximum HPS throughput you are spamming HL11. In that situation most of the time someone dies it is because your heal didn't get there in time. Say it did land in time but the target still died, it is not likely it would be different it it healed for 100 more. You would have to be lucky and have the numbers work out well for the extra 100 healed to keep the target alive. While in a lot more circumstances if your heal landed and the target still died, he wouldn't have if the heal did crit. So I would say the increased chance of the heal critting and saving the target is just as (if not more) likely then the target not dying because your minimum healed was for 100 more.
If it's possible for you to get into a scenario where a non-crit will get someone killed, he's going to die at some point no matter how much crit you have. Even with 90% crit 10% of those scenarios will end with a dead player and a wipe. Scenarios where a crit is required to save someone need to be handled in a better way than just stacking crit, while having higher +healing will completely remove some of those possible scenarios reducing the number of scenarios you have to handle differently (mostly adding more healers which will be an overkill to efficiency which you have enough if you're looking at burst, so needs to be avoided when possible, but required if you can't handle the bursts reliably). Higher crit doesn't do that.
In other words, there's a level of +healing that would elimilate deaths every time, except those that were impossible to handle in the firstplace and should've been handled by a better healing strategy. Crit will never eliminate any of them unless you stack it all the way to 100%.
Just a bit offtopic, im attempting first bosses of SSC, TK an Magtheridon. Im stacking +healing and mp5 and a bit of crit so i can get at least to 19% crit rating unbuffed. With that in mind i get 1900-2000 healing and 114 mp5 unbuffed. But now i see the top paladins in my server in ther armory with yeah, lots of healing (2400+) and crit (27-30%) BUT less mp5 than i thought (70-80 mp5). How come they need so few? is that because the use flask/consumables of mp5? I really dont get it. Can you explain me when do you change from mp5 stacking to crit stacking?
P.D.: Yeah, i have read the whole thread this past months and still feel curious why top class paladins stack crit so much.
Read the thread again. It's very easy to pick wrong stats and not be aware of it as you're probably not going to notice the difference. And on some items the crit is just on them and you don't have a choice and they're still an upgrade due to their high item level. On top of it people often pick easy upgrades even if they're not the best - be it cheap DKP or just something that is a drop and doesn't require them to farm badges (for example the boots and bracers).
As for the low mp5 it's probably becuase they don't really need more efficiency due to content on farm / shadow priest availability / wanting to maximize their burst, in which case it's all about haste and +healing. Remember a shadow priest is 200-300 mp5 and a shaman is another 62.5 mp5 + 40mp5 from mana tide with a 10k base mana pool. Those make a very significant difference to how valueable mp5 is to you (although it always will beat crit for efficiency no matter how much your party is stacked since they just don't reduce its value low enough, but they can make it worth less than +healing even from a pure efficiency standpoint). At the end you need to ask yourself "what do I need more, efficiency or burst? How much of my lasting power (efficiency) am I willing to give up for an ability to burst-heal a tank/player that took a big hit to increase my efficiency?"
Crit's not valuable for surviving worst case scenarios. But that doesn't mean that crit has no value.
Nobody said crit has no value. It has its place for efficiency and even the pure "average HPS" portion of it, although quite small, has an increase to efficiency due to allowing use of more efficient spells. But when you don't care about mana/efficiency you only care about your ability to save someone from a bad situation, in which case crit does absolutely nothing.
Just like you said, it's not going to save the day but it'll allow you to be more efficient.
Crit is not going to allow you to put less healers on the tank, as if the boss has 3000 DPS burst potential the healers on the tank need to have a 3000 HPS burst potential, without crits. They don't have to put up constant 3000 HPS, though, most of the time 2000 will be just fine so their mana doesn't have to handle constant 3000 HPS. But their non-crit burst needs to be at least 3000 HPS. Maybe sometimes they won't have to use it because a crit landed, but that doesn't reduce the amount of non-crit HPS that they need to be able to put up.
If it's possible for you to get into a scenario where a non-crit will get someone killed, he's going to die at some point no matter how much crit you have. Even with 90% crit 10% of those scenarios will end with a dead player and a wipe. Scenarios where a crit is required to save someone need to be handled in a better way than just stacking crit, while having higher +healing will completely remove some of those possible scenarios reducing the number of scenarios you have to handle differently (mostly adding more healers which will be an overkill to efficiency which you have enough if you're looking at burst, so needs to be avoided when possible, but required if you can't handle the bursts reliably). Higher crit doesn't do that.
In other words, there's a level of +healing that would elimilate deaths every time, except those that were impossible to handle in the firstplace and should've been handled by a better healing strategy. Crit will never eliminate any of them unless you stack it all the way to 100%.
I never said you should stack crit. I said you shouldn't ignore crit in burst hps.
Your logic works the same way, if at some point your heal healing for 100 more prevents a tank death you need a better healing strategy. You are looking at this way too precisely, healing isn't that exact. It is more of an art, while dps is a science. You shouldn't rely on crits, but bad luck happens to a tank. Crits can be your good luck to counter that. Say if your tank was getting some unlucky avoidance while you were spamming some medium strength heals. If a few of those crit you prolly wouldn't need to switch to bigger heal. You can't just consider worst case when you can easily have crit rates from 35-40%, and you cast multiple heals before a tank normally dies.
Sure there is a level of +heal that will eliminate burst death, but it will almost never matter how you gear yourself. The numbers would have to work out extremely well so that socketing for 22 heal instead of 11 heal 2mp5 or 10 crit will make that difference. If a tank dies from burst even though my max rank Holy Light hit him and say I didn't socket for max healing, it is extremely unlikely that if I gained 200 healing he wouldn't have died.
I wouldn't stack crit, probably not even put a crit gem in. But it is very good to have on your gear, with how diminishing returns on stacking stats on items works.
Originally Posted by galzohar
Crit is not going to allow you to put less healers on the tank, as if the boss has 3000 DPS burst potential the healers on the tank need to have a 3000 HPS burst potential, without crits. They don't have to put up constant 3000 HPS, though, most of the time 2000 will be just fine so their mana doesn't have to handle constant 3000 HPS. But their non-crit burst needs to be at least 3000 HPS. Maybe sometimes they won't have to use it because a crit landed, but that doesn't reduce the amount of non-crit HPS that they need to be able to put up.
I do not agree with this at all.
Do you ever assign healers based on average and burst dps a boss can do compared to worst case burst of the healer?
Say if I had 2500 worst case burst healing, and 3000 average burst healing. That will be fine to heal that tank in that situation. The tank isn't going to die because I did 500 less hps for like 3 or 4 seconds until one of my spells crit. In that time he is also extremely likely to avoid an attack and make the boss dps go down.
Generally speaking worst case hps is not going to allow you to put less healer on the tanks either, efficiency will. If you have any reasonable gear for the content you are doing you will be able to provide enough hps to keep the tank alive. The problem comes when you won't be able to sustain the healing cycle that is required for long enough. Sure there are some exceptions, like Brutallus, that come to mind. But in almost all those circumstances healing for 100 more worst case isn't going to help, you need to find a better healing strategy.
You aren't going to assign enough healers to tanks on Brut to heal through stomp based worst case hps and worst cast incoming damage, it is extremely inefficient. You are pretty screwed if during stomp tank dodges nothing and nothing crits. Sure you could have enough healers tank healing to prevent that. But with tanks and raid/burn healing you won't have enough dps members in the raid to kill him in time.
Actually that 2500 HPS is what, 1000 less health healed (2s cast)? That can definitely make a difference. And that 2500 HPS can last a lot longer than 3-4 seconds, just like a 60% avoidance tank is capable of not avoiding quite a few hits in a row with a higher chance to avoid than you have to crit. It's far from impossible and it is going to happen and in that case crit will not help and +healing will.
Sure there is a level of +heal that will eliminate burst death, but it will almost never matter how you gear yourself. The numbers would have to work out extremely well so that socketing for 22 heal instead of 11 heal 2mp5 or 10 crit will make that difference.
That's why you can't prove any theory, which in other words isn't helping your claim that crit helps burst. Obviously this is all talking about small differences, but the fact is out of those small differences, crit does nothing to your burst while +healing does. The only way to really understand this is looking at really big numbers: If you gained 20% crit, you would be able to heal a lot more with the same stats over a fight. But there is still an equal amount of bad scenarios where you have a chance to fail. So if you don't need any efficiency at all (and of course we always need some efficiency, but just to make a point), any amount of crit wasn't worth not gaining +healing in its place. That 20% crit chance would cost you 972 +healing (!!) - you can't tell me you'd take 20% crit over even 500 healing, now would you? (unless you cared only about efficiency and not about burst). I hope this extreme example will be enough to show how +healing helps keeping the tank alive through burst, while crit only helps efficiency and not the bad scenarios.
If you look at the actual DPS bosses are capable of putting up, you'll be surprised how much more burst HPS the healers assigned to healing him have compared to that max boss DPS. This is because putting up that max burst takes reaction time etc and nobody has the mana to actually spam his max HPS all the time, so making a boss that deals DPS anywhere near the max burst HPS of the healers would be quite impossible unless the fight was very very short.
On brutallus the tanks pretty much assume they don't get a worst case scenario and die if they do, since the chance for enough hits in a row to actually land to be unhealable is extremely low (due to the high number of hits). And once you assume a reasonable avoidance level, you go back to the "crit won't help but +healing will" - again assuming burst is your only issue and efficiency isn't.
As a side note, crit is quite inferior to other stats when it comes to efficiency as well even taking its extra HPS into account. For crit to come anywhere near other stats you need to assume you pretty much land all your heals on targets that would be able to accept a crit, which is obviously never the case. And even then it wouldn't be an incredible stat, maybe just slightly better.
One interesting project would be to parse a large number of combat logs to see how often a heal crit prevented a tank death. It'd be fairly easy to write -- you just need to track the tank's health over the course of the fight and see how often it goes under zero if the last heal crit wasn't a crit. This wouldn't be perfect, but it'd give us a much better starting point for discussing the value of crit in preventing tank deaths than the empty rhetoric we have now. If it turns out that only .01% of crits prevented a tank death, then we could just stop caring about it, and if it's a higher number you could try to evaluate the wipe preventing benefits of +healing in a similar way. Right now we don't even have bad napkin math to compare them.
Nobody said crit has no value. It has its place for efficiency and even the pure "average HPS" portion of it, although quite small, has an increase to efficiency due to allowing use of more efficient spells. But when you don't care about mana/efficiency you only care about your ability to save someone from a bad situation, in which case crit does absolutely nothing.
I think you missed the point of my entire post, or declined to read it, since you're trying to argue a point that I agree with.
Crit does not save someone from a bad situation. Instead, it makes bad situations less likely to occur. It makes it less likely that someone will have to blow a cooldown.
Crit has a more nebulous bonus to survivability because you less frequently have to recognize the fact that the tank is being underhealed and take corrective action, such as upranking, or hoping another healer switches or uses a cooldown.
These are not numbers that are simply calculated, you can't reasonably put them into a spreadsheet. But more spell crit reduces the amount of "bad luck" that you have.
I would never disagree that relying on spell crit over straight healing is foolish. But once you get to a point where you can sustain the "average" healing output required, spell crit is what is going to make things more stable, by covering those gaps in healing that inevitably appear.
Edit:
Crit is not going to allow you to put less healers on the tank, as if the boss has 3000 DPS burst potential the healers on the tank need to have a 3000 HPS burst potential, without crits. They don't have to put up constant 3000 HPS, though, most of the time 2000 will be just fine so their mana doesn't have to handle constant 3000 HPS. But their non-crit burst needs to be at least 3000 HPS. Maybe sometimes they won't have to use it because a crit landed, but that doesn't reduce the amount of non-crit HPS that they need to be able to put up.
Likewise, I never said it's going to allow you to put signifigantly less healers on the tank.
Lets take Brutallus for an example, because it's one of the first bosses with non-trivial healing allocation. (Pretty much every other fight in the game, save maybe Bloodboil is a joke in terms of healing in my opinion.)
Imagine if he could not miss any of his attacks or have them dodged/parried. You would not at all be able to keep up the tank. Especially in this fight, healers are a scarce resource. You can not afford to stack healers because then you will not make the DPS requirement.
You're relying on averages, you're relying on chance. You're tuning your healers to what they most frequently need to heal. Now again, I never said that you are relying on crits to save the tank. You're relying on crits to more frequently put you in a situation where the tank doesn't need any special interaction to be saved. Not completely remove the chance that the tank must be saved, but to lower the number of times it's an issue.
If your tank does have an unlucky avoidance streak, there is a chance you have a few crits. This buoy's the tank to a position where you don't require an NS, or a last stand, or some other cooldown. The more frequently this happens, the less frequently a cooldown or corrective action is required. The less frequently corrective action is required, the less likely that it will be unavailable, or too slow.
The number of times a crit actually saves the tank (A crit heal lands healing him to the point where the next hit would kill him were the heal a normal one) is completely irrelevant. Because the benefit of the crit healing is that those sorts of situations, where the tank is in imminent risk of dying, are less likely to ever arise in the first place with a higher crit rate.
Actually that 2500 HPS is what, 1000 less health healed (2s cast)? That can definitely make a difference. And that 2500 HPS can last a lot longer than 3-4 seconds, just like a 60% avoidance tank is capable of not avoiding quite a few hits in a row with a higher chance to avoid than you have to crit. It's far from impossible and it is going to happen and in that case crit will not help and +healing will.
That's why you can't prove any theory, which in other words isn't helping your claim that crit helps burst. Obviously this is all talking about small differences, but the fact is out of those small differences, crit does nothing to your burst while +healing does. The only way to really understand this is looking at really big numbers: If you gained 20% crit, you would be able to heal a lot more with the same stats over a fight. But there is still an equal amount of bad scenarios where you have a chance to fail. So if you don't need any efficiency at all (and of course we always need some efficiency, but just to make a point), any amount of crit wasn't worth not gaining +healing in its place. That 20% crit chance would cost you 972 +healing (!!) - you can't tell me you'd take 20% crit over even 500 healing, now would you? (unless you cared only about efficiency and not about burst). I hope this extreme example will be enough to show how +healing helps keeping the tank alive through burst, while crit only helps efficiency and not the bad scenarios.
If you look at the actual DPS bosses are capable of putting up, you'll be surprised how much more burst HPS the healers assigned to healing him have compared to that max boss DPS. This is because putting up that max burst takes reaction time etc and nobody has the mana to actually spam his max HPS all the time, so making a boss that deals DPS anywhere near the max burst HPS of the healers would be quite impossible unless the fight was very very short.
On brutallus the tanks pretty much assume they don't get a worst case scenario and die if they do, since the chance for enough hits in a row to actually land to be unhealable is extremely low (due to the high number of hits). And once you assume a reasonable avoidance level, you go back to the "crit won't help but +healing will" - again assuming burst is your only issue and efficiency isn't.
As a side note, crit is quite inferior to other stats when it comes to efficiency as well even taking its extra HPS into account. For crit to come anywhere near other stats you need to assume you pretty much land all your heals on targets that would be able to accept a crit, which is obviously never the case. And even then it wouldn't be an incredible stat, maybe just slightly better.
What I am saying is you can't ignore it when it comes to burst healing, and you seem to think I am advocating stacking it to keep tanks alive.
You can't just ignore efficiency. Sure if we didn't have illumination I would not gear for crit at all. But we do, and efficiency matters. Crit is balanced stat for both hps and efficiency. Yes, crit isn't good at it as healing/mp5 at doing it. But with how items work with diminishing returns it is good to get on your gear, like T6 boots/belt are perfectly itemized.
That is an unrealistic example. We can't make choices on that scale, we have to make it on the margin. Especially since how much crit improves your hps improves with your +healing, crit is better to have when you have 2500 +healing then when you had 1500 +healing. Show me some gear set that has 972 more healing then one with 20% more crit.
And yes, I would take 20% crit over 500 +healing. 500 healing increases your holy light by 350 healing. 20% crit would provide and average increase of 500 (presuming 5000 base HL), and reduces the average cost of your healing spell by 12%! I would definitely take 500 average increased heal and 12% average spell cost reduction over 350 worst case healing. Much more frequently someone dies because I am trying to conserve my mana because it is low the max rank Holy Light isn't enough. Ideally though I would take a balance of the two.
How often does a tank die while spamming HL11? Almost never in my experience.
I love +healing and it works very well for paladins. But this isn't because I have better worst cast burst heals. It is because Flash of Light has an amazing coefficient for how much mana it costs. So it scales amazingly well with +healing.
One interesting project would be to parse a large number of combat logs to see how often a heal crit prevented a tank death. It'd be fairly easy to write -- you just need to track the tank's health over the course of the fight and see how often it goes under zero if the last heal crit wasn't a crit. This wouldn't be perfect, but it'd give us a much better starting point for discussing the value of crit in preventing tank deaths than the empty rhetoric we have now. If it turns out that only .01% of crits prevented a tank death, then we could just stop caring about it, and if it's a higher number you could try to evaluate the wipe preventing benefits of +healing in a similar way. Right now we don't even have bad napkin math to compare them.
It really isn't so simple. I totally agree with zeidrich views on crit. How many times have you said "thank god that crit", or at least thought something like that. It might not have been something so direct like if it didn't crit he would die, but it got the tank much closer to being topped off. Heals very much have a butterfly effect, that whole thing of a butterfly 1000 miles away causing a torando. You can't just look at the last heal, it is a sequence of events that makes you decide what heals to cast. Say if a tank got low, like down 10-12k health, you cast a max rank Holy Light. It doesn't crit, and tank is still decently low so you cast another (maybe a slightly lower rank like 9). If it does crit the tank is only down 2-3k health, so back to FoL spamming.
It really isn't so simple. I totally agree with zeidrich views on crit. How many times have you said "thank god that crit", or at least thought something like that. It might not have been something so direct like if it didn't crit he would die, but it got the tank much closer to being topped off. Heals very much have a butterfly effect, that whole thing of a butterfly 1000 miles away causing a torando. You can't just look at the last heal, it is a sequence of events that makes you decide what heals to cast. Say if a tank got low, like down 10-12k health, you cast a max rank Holy Light. It doesn't crit, and tank is still decently low so you cast another (maybe a slightly lower rank like 9). If it does crit the tank is only down 2-3k health, so back to FoL spamming.
There's three effects a crit can have:
1) Save mana, either directly via Illumination or indirectly via reducing the amount of healing you have to do in the future
2) Influence your future casts
3) Directly prevent a tank death sometime between when the heal lands and when the total overhealing on the tank is equal to the size of the crit bonus (which could be well more than one spell, so only looking at the last crit would be wrong)
1 is easy to model, and has been modeled many times. For 3, you don't need to look at a very large section of the log for each crit heal. If your base HL heals for 6k, then once the tank has been overhealed for 3k following a HL crit, if the tank never dropped under 3k health then the only the crit had was saving you (or other healers) mana, as that overhealing would have covered the damage your crit did. If the tank does dip under 3k health, then the crit prevented a wipe and was more useful than predicted by the efficiency/average hp/s model.
Your point seems to be that #2 is too significant for anything that neglects it to be useful. If you're solo-healing a tank (or doing a 5-man/etc.), it probably is too complex to model. However, in the situation of several healers on a tank, it isn't. There's really only two things it can do: make you cast too small of a heal next, leading to tank death (or forcing another healer to cover for your mistake, potentially causing a tank death in the future), or it can uneventfully save you some mana. We don't need to do anything special for the second case, as any decent efficiency model will already reflect your actual spell usage. If the first case happens with any frequency... you're a bad player, and should stop healing reactively in situations where it isn't safe.
Looking only at the times when a crit directly prevented a death in the short-term doesn't come anywhere close to fully modeling the value of crit, but that doesn't matter -- all it has to do is model the part of the value that isn't already modeled.
Remember I'm always seperating the value of efficiency and HPS. If you have people die because you're conserving mana, then what you need is obviuosly efficiency over burst HPS, in which case healing/mp5 would be best anyway.
For burst and saving a tank right before he dies, crit is like avoidance -has the chance to help but is nothing reliable so it doesn't count towards your ability to reliably saving him. Burst is far from the only thing - you can't burst heal with no mana etc. But some fights you'll be able (and want) to neglect other things for burst and other fights you won't.
For efficiency, or "doing more healing using all my mana", crit definitely helps. Based on my calculations it's somewhat weaker than healing/mp5 but a bit stronger than haste, and it only gets anywhere near healing/mp5 if you assume your crits never overheal (and I'm not even talking about causing other healers to overheal... simplifying it to assume that if you didn't overheal you didn't cause overhealing which is actually highballing the value of crit), meaning you have to have all your 5k (7.5k crit) HLs to land on a target that can accept a 7.5k heal with no exceptions. In reality this is far from the case and therefore crit loses value far under other stats. If you counted all crits as full heals it would still only be more or less on par with other stats in terms of efficiency.
It's not practical to mix the effects of burst and efficiency, as the one you actually need more will depend on the fight and raid/party composition. But calculating how much stats help either of those, seperately, is very possible, and leaving it up to you to decide which of those is more important and by how much - or in other words how much efficiency you're willing to give up for how much burst HPS.
Basically I am looking for some insight into when, how and where you Judge. I'm no Healadin, but whenever I try to bring the subject up in my guild I get the feeling that I'm pushing the healadins too far. Personally I think the issue lies with the guild not having raided with a Retribution Paladin for ages, if ever before now. So the Healadins are either quite unaware of how it works or are simply way out of shape for that.
So on their behalf I would like a few pointers.
Do you run in with the tank and judge before he takes too much damage, while having the other healers take up the slack for a few seconds? Or do you seal up and heal your way closer until you are close enough for a Judgement?
Basically what are your tactics for this?
Being a non-Paladin I don't think it is my place to give them tactical advice, especially since my own Paladin is merely level 26.
If you find it easy to do, then I might have a bit of ammo to pound them with.
For efficiency, or "doing more healing using all my mana", crit definitely helps. Based on my calculations it's somewhat weaker than healing/mp5 but a bit stronger than haste, and it only gets anywhere near healing/mp5 if you assume your crits never overheal (and I'm not even talking about causing other healers to overheal... simplifying it to assume that if you didn't overheal you didn't cause overhealing which is actually highballing the value of crit), meaning you have to have all your 5k (7.5k crit) HLs to land on a target that can accept a 7.5k heal with no exceptions. In reality this is far from the case and therefore crit loses value far under other stats. If you counted all crits as full heals it would still only be more or less on par with other stats in terms of efficiency.
That isn't really true. That only holds if all of your non crit heals don't overheal at all. Need to compare the amount overhealed on crits versus non crit.
Can you show some numbers to back up crit heals overhealing more compared to normal heals.
There's three effects a crit can have:
1) Save mana, either directly via Illumination or indirectly via reducing the amount of healing you have to do in the future
2) Influence your future casts
3) Directly prevent a tank death sometime between when the heal lands and when the total overhealing on the tank is equal to the size of the crit bonus (which could be well more than one spell, so only looking at the last crit would be wrong)
1 is easy to model, and has been modeled many times. For 3, you don't need to look at a very large section of the log for each crit heal. If your base HL heals for 6k, then once the tank has been overhealed for 3k following a HL crit, if the tank never dropped under 3k health then the only the crit had was saving you (or other healers) mana, as that overhealing would have covered the damage your crit did. If the tank does dip under 3k health, then the crit prevented a wipe and was more useful than predicted by the efficiency/average hp/s model
Has someone modeled how a heal criting (or even just bigger) can reduce the amount (and therefor mana cost) spent to heal in the future? I haven't seen one, and to me it seems very difficult to model. There are a lot of things that are very circumstantial that you can't really model. It depends on the tanks health relative to how much damage he is taking how frequently.
Say on Anetheron if my tank gets low for some reason (lot of healers carrion swarmed for example) I am much more likely to cast Flash of Light after my Holy Light because Anetheron does not really do that much burst to a tank. But on a fight like Azgalor where he can do a lot more burst with his heavy melee attack and cleave, especially if his Silence CD is up I will be much more prone to Holy Light even though it might not be needed. Even though in both situations they would be down the same hp, would essentially need a model for each boss.
Basically I am looking for some insight into when, how and where you Judge. I'm no Healadin, but whenever I try to bring the subject up in my guild I get the feeling that I'm pushing the healadins too far. Personally I think the issue lies with the guild not having raided with a Retribution Paladin for ages, if ever before now. So the Healadins are either quite unaware of how it works or are simply way out of shape for that.
So on their behalf I would like a few pointers.
Do you run in with the tank and judge before he takes too much damage, while having the other healers take up the slack for a few seconds? Or do you seal up and heal your way closer until you are close enough for a Judgement?
Basically what are your tactics for this?
Being a non-Paladin I don't think it is my place to give them tactical advice, especially since my own Paladin is merely level 26.
If you find it easy to do, then I might have a bit of ammo to pound them with.
Depends alot on fight. In general I feel that there is so much movement going on at the start of a fight that I often wait for other people to get in place to start moving myself. That is a great moment for a judgement.
The problem after that is that once you reach t6, fights get more and more healing intensive and you might just not have the global cooldown + movement time to spare to judge. But on fights where you are in range of the boss (teron, naj'entus possibly, RoS, etc) and loosing 1 GC wont kill someone, throwing a judgement every now and then is not too hard.
If you are still learning fights though, it does take quite alot of focus that you could be spending on healing/positioning. If you really want judgements to be up, get a ret pala
Now I read your question again though, I assume you have a ret pala and your holy paladins are not capable of putting up just 1 judgement at the start and possibly another one when your ret pala informs them judgements droppe? Thats pretty much slacking.
(Judgements do not have debuff priority though so be sure to monitor if they aren't being kicked off)
A lot of our boss pulls are misdirect pulls and so if your paladins position themselves between the boss and the tank they can judge as he goes passed. If its not a misdirect pull I run up with the tank and judge as soon as he gets aggro.
We run with 1 Retri and hes not in every raid so at times he will occaisionally remind us to judge if we have got used to not having him in for a bit.
If your Holy Paladins feel they can't get one judgement off at the start of the fight because the tank might die then theres something wrong in my opinion as with a little planning its not a difficult thing to do.
So getting the judgement in early seems to be the easiest for most fights? But in some cases waiting until the fight is 'stable' is better. Hm, pretty much as I thought, but I really wanted some authority on this. Since I have no experience as a Pally raider I really didn't have any idea of how hard or pressing it might be.
Btw, are are on mid-T5. I think it is 2-TK and 4-SSC. So I don't know if the bosses are as healingintensive as you experience on T6.
Judging it early is benefical to save the gcd, but you should tell them to be aware of the situation and not get tunnel visioned. It should be noted that the bosses hit for the most at the start of the pull when demo/tclap are not up yet. Keeping it up with a ret paladin should be fine, it's a lot easier to just judge once and forget than to keep reapplying every 20 seconds with a swing or a rejudge. It's also I think more beneficial later into the fight than earlier, because then you know mana for DPS and whatnot is not full/on cooldown.
On another topic, not to rehash the crit vs healing argument, I'm up to 118 spell-haste from re-gemming yellow sockets to spell-haste and I'm liking it a lot, more so than when I had crit gems in the same yellow sockets. 1.4 speed FoL before a scarab proc is pretty nice, and its 1.8 or so HL after LG. And not to break Gurgthock's rule about posting Sunwell strats, spiritual attunement really really shines in Sunwell for longevity, and combined with a spriest it's a ton of mana return allowing you to spam the neccesary HLs for the tank to survive. I don't want to go any deeper for fear of breaking the rule, but I really like the synergy of spiritual attunement now, and unfortunately I forget when it was added to paladins =/
The main reason tanks die on pulls is less that they hit a bit harder/faster, more that the healers are simply not ready and/or not precasting heals. Think of all the times you only started healing 5 seconds into the fight for some reason, however rare it may had been. Now think if all healers just happened to do it for some reason on the same pull => dead tank. That causes most healers to be on their toes on pulls so that when 1-2 healers mess up the tank still doesn't die, but it only means that if all healers focus you can easily put up that judgement on the pull. In other words bosses don't deal enough dps to kill a tank when all healers in the raid are available to heal him. It's just too much healing on too little DPS, some/most healers simply need to not heal (or sometimes just not pre-casting) for the tank to die.
As for crit overhealing, while I hand't found a good way to model it, the effective healing you get from a crit on average is definitely less than 50% and definitely higher than 0%, and probably not very close to either. That alone is enough to determine crit is not an awesome stat for how much it costs, even for pure efficiency where it actually does help. I gave a rough estimationof 20% based on intuition but would love to get something with more basis to it.
As for crit overhealing, while I hand't found a good way to model it, the effective healing you get from a crit on average is definitely less than 50% and definitely higher than 0%, and probably not very close to either. That alone is enough to determine crit is not an awesome stat for how much it costs, even for pure efficiency where it actually does help. I gave a rough estimationof 20% based on intuition but would love to get something with more basis to it.
Can you show anything to back that up? Besides just bolding it and stating it as true. How is that enough to show it isn't worth it? You need to show how much extra crit adds to healing (excluding overheals) versus how much extra and equivalent amount of +healing would add (excluding overheals) while considering how they effect your mana usage.
I agree crit isn't the best stat to stack for gems and such, but it is definitely good enough to look for balanced gear with it. You can't just ignore it in your discussion of stats which you do sometimes.
04/07/08, 5:43 AM
Similar Threads
