It Gets Better: A discussion of healing metrics and alacrity

[Daross]

Such a great and long post with lots of math, but still miss small detail.

That even with 0 altracity, constant casting spells spent more energy, than you regain...

Which means you cast ability, then pause (pause is adequate to amaunt of altracity), then cast again.

No matter how much altracity you have, you do roughly same amount of HPS (with casting abilities).

Ofc, thanks to lower downtime, you have more time to move, or cast no-cost abilities. Since no-cost abilities heals very little its value is more as utility.

 

So unless you are Sage or Sorc, altracity IS NOT HPS stat.. its utility.

Helps your mobility, support triage, helps getting from low energy situations... but does not affect HPS directly.

[ironix]

I appreciate your work on this! Great job. =D

 

However, it just occurred to me. Why bother doing all this work in .NET? Wouldn't a tool like this be better served as a web application? You could probably do it up in Ruby pretty quickly. Hell, a site like torhead or the SWTOR equivalent to maxdps would probably love to have a tool like this on their site.

 

I considered doing one up based on your code, but alas, it has very few comments. =(

[ZraveX]

I have not analyzed your code or what your sim is doing in detail, but I have some suggestions/comments.

 

1) In my opinion, the "wiggle" of the graphs greatly reduces their effectiveness, just from a statistical perspective (large error). I suggest you run the simulation multiple times with different initial (or random) parameters and then average the result, that should take care of it.

 

2) I believe that most of your conclusions are derived from a single fact - that surge becomes worthless very quickly. The higher your gear level, the higher the initial surge, and the sooner it reaches devastating DR. Considering that the only alternative is alacrity, the potential is obvious.

 

This is illustrated in the +/- 200 graphs where in all 3 cases surge is the one you can least afford to lose, while alacrity is the one that least reduces your effectiveness when lost.

 

3) Your second graph actually shows that alacrity is the worst stat. First there is a problem with the plots: they all start from arbitrary y-intercepts related to how much of that stat the tier sets have originally. Comparing the different heights of the plots is entirely meaningless.

 

To illustrate the problem, at the 500 x-coordinate you are comparing which of the following gains is better:

 

a) +312 Power

b) +144 crit

c) +296 surge

d) +194 alacrity

 

Which of course produces no useful result. I suggest you do the simulations with all 4 stats starting at 250, at least for the second type of plot.

 

What we can compare, however, is the slope of the graphs. The incremental benefit of adding one more point in one of the 4 stats can be approximated by the total increase in y for each plot. The domain to use is tricky however. Using the full plot doesn't work because surge is basically a flat line after 500, and honestly I'm not sure if you can reach that high without using augments, which is silly. If you average over the 0-500 domain, I'm eyeballing the following increases:

 

Alacrity: ~70

Power: ~150

Surge: ~110

Crit: ~140

 

This is very rough of course, compounded by the unequal total stat allocation. But it hints that for realistic maximum values, alacrity is overall the worst stat. It should be, however, better than the last points of surge due to the crushing DR.

 

If you can do a derivative plot of the stats (after fixing the unequal allocation and smoothing), the point where dHPS/dsurge crosses dHPS/dalacrity will tell us when alacrity becomes better. This would be a very useful result!

Edited March 10, 2012 by ZraveX

[ErrantMercenary]

Such a great and long post with lots of math, but still miss small detail.

That even with 0 altracity, constant casting spells spent more energy, than you regain...

Which means you cast ability, then pause (pause is adequate to amaunt of altracity), then cast again.

No matter how much altracity you have, you do roughly same amount of HPS (with casting abilities).

Ofc, thanks to lower downtime, you have more time to move, or cast no-cost abilities. Since no-cost abilities heals very little its value is more as utility.

 

So unless you are Sage or Sorc, altracity IS NOT HPS stat.. its utility.

Helps your mobility, support triage, helps getting from low energy situations... but does not affect HPS directly.

 

 

Uhm, did you read the post? This is quite literally the point of the new analysis. The old model really only accounted for Alacrity as a passive benefit stat instead of one that required a bit of active thought (and modeling) to account for it's utility. However, it is worth mentioning that Alacrity is still a stat that offers passive benefit for us Mandos and Mercs because especially at the higher values because when paired with the PvE set bonus (+3 seconds on SCC/SCG) we can actually achieve some measure of the more WoW-esque model where we put more healing on the board by cramming more "good" spells into the same timeframe in emergency situations. Especially in the raid healing scenarios where our AoE healing is so lacking the smart decision is generally to bull rush more AP/MP casts on individual targets in dire need.

[RuQu]

I have not analyzed your code or what your sim is doing in detail, but I have some suggestions/comments.

 

1) In my opinion, the "wiggle" of the graphs greatly reduces their effectiveness, just from a statistical perspective (large error). I suggest you run the simulation multiple times with different initial (or random) parameters and then average the result, that should take care of it.

 

Unfortunately, there is 0 RNG in Combat Medic healing. Repeated runs will produce identical results. I do, of course, run multiple iterations and average for Gunnery DPS and Scoundrel Healing because those involve random procs. Since the only RNG at all present in the CM rotation is the standard min-max ranges on heals and crit chance, both of those are easily account for using the Expected Healing Value.

 

The wiggle comes from a couple of sources, but primarily it is just changes in the rotation made based on the data available at the time of each decision.

 

I think part of this comes from using unrealistically small changes in stat values. Swapping out mods should produce a minimum change of 30, and I am using a change of 5. As such, I am oversampling, increasing the "wiggle" and increasing the computation time.

 

2) I believe that most of your conclusions are derived from a single fact - that surge becomes worthless very quickly. The higher your gear level, the higher the initial surge, and the sooner it reaches devastating DR. Considering that the only alternative is alacrity, the potential is obvious.

 

This is illustrated in the +/- 200 graphs where in all 3 cases surge is the one you can least afford to lose, while alacrity is the one that least reduces your effectiveness when lost.

 

This is correct.

 

3) Your second graph actually shows that alacrity is the worst stat. First there is a problem with the plots: they all start from arbitrary y-intercepts related to how much of that stat the tier sets have originally. Comparing the different heights of the plots is entirely meaningless.

 

Not sure I follow you. Are we still talking about the second graph? On the second graph of each pair, the values are 0-1000 for the stat specified on the curve. The different heights on the plot should give you a sense of the over-all value of the stat at that gear level (for instance Alacrity is the worst stat in Tionese), but gearing changes should look more to the first graph.

 

To illustrate the problem, at the 500 x-coordinate you are comparing which of the following gains is better:

 

a) +312 Power

b) +144 crit

c) +296 surge

d) +194 alacrity

 

Which of course produces no useful result. I suggest you do the simulations with all 4 stats starting at 250, at least for the second type of plot.

 

They all start at 0.

 

At the 500 x-coordinate they are all 500 on their respective curve.

 

Although I think I might see what you are getting at...a rating of 500 is not a fixed gain of rating value. If you want to see the gains you would have to compare Alacrity at 500 to Surge at 398. This is true...but it is also why the first graph exists to do exactly that. The second graph largely puts the DR in perspective and I don't honestly think it is useful for much more than a "huh, that's interesting."

 

What we can compare, however, is the slope of the graphs. The incremental benefit of adding one more point in one of the 4 stats can be approximated by the total increase in y for each plot. The domain to use is tricky however. Using the full plot doesn't work because surge is basically a flat line after 500, and honestly I'm not sure if you can reach that high without using augments, which is silly. If you average over the 0-500 domain, I'm eyeballing the following increases:

 

Alacrity: ~70

Power: ~150

Surge: ~110

Crit: ~140

 

Why not use the first graph and pull the local slop off directly?

 

This is very rough of course, compounded by the unequal total stat allocation. But it hints that for realistic maximum values, alacrity is overall the worst stat. It should be, however, better than the last points of surge due to the crushing DR.

 

If you can do a derivative plot of the stats (after fixing the unequal allocation and smoothing), the point where dHPS/dsurge crosses dHPS/dalacrity will tell us when alacrity becomes better. This would be a very useful result!

 

I'm working with some people to come up with a way to efficiently compute (in CPU time) the points where those cross, and how best to present that data.

Edited March 10, 2012 by RuQu

[RuQu]

Oops....my mistake.

 

I forgot I had re-turned on including the HS cast time, so I was plotting HPCT not HPRCT.

 

The result was that wiggle. When you decide "Meh, I don't really need to cast a heal right now, let me just top him off with HS," the result is a drop in HPCT because HS is so very weak but takes 1.5s to cast.

 

Graph Showing Change in HPRCT when changing a stat -200 to +200

 

Here is the Rakata Stat Curve of HPRCT, properly excluding HS.

 

I'll update the OP here once I rerun all three.

Edited March 10, 2012 by RuQu

[RuQu]

I appreciate your work on this! Great job. =D

 

However, it just occurred to me. Why bother doing all this work in .NET? Wouldn't a tool like this be better served as a web application? You could probably do it up in Ruby pretty quickly. Hell, a site like torhead or the SWTOR equivalent to maxdps would probably love to have a tool like this on their site.

 

I considered doing one up based on your code, but alas, it has very few comments. =(

 

It's not .NET, it's C++ and Qt.

 

The primary reason: I don't know Ruby.

 

Honestly, my background is in light Matlab programming for tidal and geodetic analysis. I needed to learn C/C++ for work, and figured I'd make more progress learning to code if I coded something I cared about...which meant making a tool to help me figure out defensive stats for my Shadow and healing stats for my Commando. I already had spreadsheets, this just opened up the options.

 

Then...no one else was doing similar work. Lots of spreadsheets around, some simulations for Sorc DPS, but that's it. So it just sort of snow-balled from there.

 

I apologize for the comments. It certainly needs more, and it needs me to clean out the outdated code that I currently have commented out because I was trying alternate methods.

 

I did some looking over of the Ruby wikipedia page. It would certainly be useful as a web-tool, but I think I first need to get the code optimized to minimize the number of iterations and computation time. As an interpreted language, calculation times will increase, and I think people might just assume the website crashed.

 

As an educational work in progress, there is certainly room to improve the efficiency of the code. Once I get it to where I can stop labeling every release as an alpha or a beta, it might be time to sit down and look at how it could be ported to a web-app.

 

In short, a very good idea, I just don't think its ready yet.

[RuQu]

There we go, all updated. I left the HPCT charts in if people are interested in them.

 

Synopsis:

 

Power > Crit at pretty much all levels.

Surge > Alacrity at low levels, mostly caused by the sharp DR curve on Surge.

Alacrity > Surge at Columi and above, essentially one you get above ~125 Surge the slope of the Alacrity curve is better, and over 200+ the Surge DR gets pretty brutal.

[ErrantMercenary]

There we go, all updated. I left the HPCT charts in if people are interested in them.

 

Synopsis:

 

Power > Crit at pretty much all levels.

Surge > Alacrity at low levels, mostly caused by the sharp DR curve on Surge.

Alacrity > Surge at Columi and above, essentially one you get above ~125 Surge the slope of the Alacrity curve is better, and over 200+ the Surge DR gets pretty brutal.

 

That was actually the most telling fact I pulled from these new graphs. We all saw the Surge changes go in, but since I was still operating on the premise that Alacrity caused trouble beyond your previously recorded sweet spots, I let it stand in (I'm in what we'll call Columi+, Columi or equivalent across the board, plus a couple pieces of Rakata scattered throughout) but they've essentially implemented a hard cap on Surge. It'll be interesting to see how they adjust this mechanic down the line so that they can still put Surge on new levels of gear as new content becomes available. As it is, it's as close to a zero value stat past the 200 marker as we're ever liable to see.

[RuQu]

That was actually the most telling fact I pulled from these new graphs. We all saw the Surge changes go in, but since I was still operating on the premise that Alacrity caused trouble beyond your previously recorded sweet spots, I let it stand in (I'm in what we'll call Columi+, Columi or equivalent across the board, plus a couple pieces of Rakata scattered throughout) but they've essentially implemented a hard cap on Surge. It'll be interesting to see how they adjust this mechanic down the line so that they can still put Surge on new levels of gear as new content becomes available. As it is, it's as close to a zero value stat past the 200 marker as we're ever liable to see.

 

Yeah, there are no "hard caps" in this game like you saw in WoW...but the DR on Surge is so bad as to be almost the same thing.

 

I've got like 15 irons in the fire right now, but I need to make some time to upgrade the CM guides accordingly.

 

Sorry again for any confusion putting up the wrong graphs before.

 

Luckily this initial round of content is easy enough that it is doable even without optimal gear setups. They claim 1.2 will be much harder, but they are also giving us more tools for confirming our models, so it should balance out.

[ZraveX]

For the smoothing, you can make the boss damage random if the healing is insufficiently so. You can also make the simulation finer and then average clumps. I noticed that your latest graphs are a lot smoother now, so this isn't as important anymore.

 

Did you account for clipping of the HS hot during supercharged gas? I find that it is sometimes relevant, but quantifying it would be tricky. During my own analysis, I set a limit of alacrity at 14% so that it couldn't ever happen, but I am in general less optimistic about alacrity than you are.

 

I agree with your assessment that suboptimal setups can still prosper in the current content. But healing requirements will go up at the next content level, which if I understand your analysis correctly will skew your conclusions. You may want to look into what happens at much greater boss damage levels.

 

Personally, I find healing this content so easy that I use Muzzle Fluting with the 2pc eliminator set bonus and dps during tank cooldowns or burst phases. This makes the "optimal" healing turret setup (i.e. the one that can heal most effortlessly) a bit of a moot point for me.

[RuQu]

For the smoothing, you can make the boss damage random if the healing is insufficiently so. You can also make the simulation finer and then average clumps. I noticed that your latest graphs are a lot smoother now, so this isn't as important anymore.

 

Did you account for clipping of the HS hot during supercharged gas? I find that it is sometimes relevant, but quantifying it would be tricky. During my own analysis, I set a limit of alacrity at 14% so that it couldn't ever happen, but I am in general less optimistic about alacrity than you are.

 

I actually was using the wrong time for the HoT. I was using the 2s reported on the darthater skill calculator, instead of the correct 9s. Fixing that is on my todo list for the week. Clipping was irrelevant at 2s since AP/MP (HSc/RSc) would never clip it. At 9s, I will have it clip correctly.

 

Even at 14% Alacrity it would still clip. If you cast AP/MP/AP that is 1.5/2/1.5s, and the HoT applies after the initial cast, so you only have 3.5s between when the HoT starts and the next AP lands. So you will only get 1 tick. I'm assuming you mean you set a limit at 14% so you always get at least 1? (3/3.5 = 0.857, or 14.3% Alacrity).

 

That is a solid recommendation. I'll note it as a consideration for people in the guide, and I'll rerun things after the HoT is fixed to see if the clipping produces a noticeable change in the alacrity curve at 14%.

 

I agree with your assessment that suboptimal setups can still prosper in the current content. But healing requirements will go up at the next content level, which if I understand your analysis correctly will skew your conclusions. You may want to look into what happens at much greater boss damage levels.

 

Personally, I find healing this content so easy that I use Muzzle Fluting with the 2pc eliminator set bonus and dps during tank cooldowns or burst phases. This makes the "optimal" healing turret setup (i.e. the one that can heal most effortlessly) a bit of a moot point for me.

 

Frankly, the boss damage numbers of the current estimate are fairly arbitrary. When we get new content we will also get combat logs and the arbitrary nature can be removed and replaced with content level correct values. I think the concept in use is valid though, but the results will surely change a bit as the model shifts to better damage values.

[XtremJedi]

I read the first page, but didn't have time to read the rest, as I have to leave soon... so forgive me if I repeat something others have said.

 

There have been several big thread discussions in this forum about the usefulness of Alacrity. Many of us pointed out and argued the basic point here; that you can have more time between casts if your alacrity is high. I'm not sure how you missed the discussions.

 

What I don't see mentioned in the OP though, is that one of the major gains from this, is that it can allow you to do other stuff too. The extra time you gain between the heals needed to counter incoming damage, can be used for cc, stuns, interrupts, debuff removers etc etc. Those themselves can lead to less incoming damage, which in return grants you even more benefit.

 

Connected with that, is that you also don't really directly mention the benefits to burst healing. If there's a period of excessive damage incoming and my Cool Head is off CD, then my Sawbones can push out a higher burst of heavy healing and other effects during that period, because of alacrity, then recover by using Cool Head to pull me back into a good regen spot.

 

It is still something that my Sorc can take more advantage of than my Sawbones, but the value is there for both.

 

X

[RuQu]

I read the first page, but didn't have time to read the rest, as I have to leave soon... so forgive me if I repeat something others have said.

 

There have been several big thread discussions in this forum about the usefulness of Alacrity. Many of us pointed out and argued the basic point here; that you can have more time between casts if your alacrity is high. I'm not sure how you missed the discussions.

 

The most commonly mentioned use of that time was more time for Hammer Shot, to which the most common response was "Yeah, but Hammer Shot is terribly weak and instant, so you would get more from Power." Trying to see just how many extra HS casts you would earn, and how that would compare to just making everything stronger, is sort of where this all started.

 

What I don't see mentioned in the OP though, is that one of the major gains from this, is that it can allow you to do other stuff too. The extra time you gain between the heals needed to counter incoming damage, can be used for cc, stuns, interrupts, debuff removers etc etc. Those themselves can lead to less incoming damage, which in return grants you even more benefit.

 

It isn't mentioned for a couple reasons. First, many of those aren't relevant on a boss fight, since they are immune. Second, they are almost impossible to put a number on. Third, it is sort of implicit that if you have more time, you can do stuff with it. Whether you DPS, CC, /clubdance, whatever is up to you, your resource level at the time, and exactly what content you are doing.

 

Connected with that, is that you also don't really directly mention the benefits to burst healing. If there's a period of excessive damage incoming and my Cool Head is off CD, then my Sawbones can push out a higher burst of heavy healing and other effects during that period, because of alacrity, then recover by using Cool Head to pull me back into a good regen spot.

 

It is still something that my Sorc can take more advantage of than my Sawbones, but the value is there for both.

 

X

 

These calculations cover a decent range of damage, from trivial to lethal. At the lethal level, RC is used when Ammo drops too low due to needing to chain cast.

 

If you want to examine burst outside of that, you need to define the problem better. We aren't discussing DPS here, so we won't intentionally blow all of our resources every 2 minutes just to use CH/RC/VH/Adrenaline Probe. To factor it in would require that there be a spike of damage so great that your resources drop below the threshold to use it, and then what happens when that happens again before 2 minutes are up? Super short fights, such as PvP aren't really within the scope of discussion. I am up to discussing the relative merits of Alacrity in terms of burst, but you need to define the conditions of burst better first. How long of a fight, how much time between damage spikes, duration of spikes, etc.

[XtremJedi]

I think that one of the great benefits of alacrity is precisely that it is particularly useful when things do go slightly wrong. DPS accidentally breaks a cc or pulls aggro, somebody stands in stuff when they shouldn't etc etc.

 

Now, it may well be true that these events don't happen all that often, so there's an argument to say that your equipping decisions shouldn't be based around those exceptional circumstances.... but there's also an argument for saying, that if you can manage the predictable situations anyway - then having the extra available for the exceptions makes it worthwhile.

 

People often say of healers other than Sorc/Sage that there is too little leeway for errors. If the group/party doesn't play perfectly, then things tend to go south more quickly. A stat that perhaps allows for better recovery when those errors occur, might well be seen as being more useful.

 

Whether you can formulate that within a mathematical model is highly doubtful. These will tend to be exceptional events. But sometimes preparing for the exceptions is a wise thing to do...

 

X

[RuQu]

I think that one of the great benefits of alacrity is precisely that it is particularly useful when things do go slightly wrong. DPS accidentally breaks a cc or pulls aggro, somebody stands in stuff when they shouldn't etc etc.

 

Now, it may well be true that these events don't happen all that often, so there's an argument to say that your equipping decisions shouldn't be based around those exceptional circumstances.... but there's also an argument for saying, that if you can manage the predictable situations anyway - then having the extra available for the exceptions makes it worthwhile.

 

People often say of healers other than Sorc/Sage that there is too little leeway for errors. If the group/party doesn't play perfectly, then things tend to go south more quickly. A stat that perhaps allows for better recovery when those errors occur, might well be seen as being more useful.

 

Whether you can formulate that within a mathematical model is highly doubtful. These will tend to be exceptional events. But sometimes preparing for the exceptions is a wise thing to do...

 

X

 

Some very good points, and I think this parallels with what was said before.

 

When you are barely able to heal the content, ie progression, everyone needs to be not making mistakes and you are on the edge of chain casting just to keep your HPS throughput matching the DPS incoming. Once your gear lets you match the DPS, you now have the the freedom to increase your speed and open up your options.

 

I think it's precisely because the healers have that extra freedom to fix other people's mistakes once the tank dying isn't a constant threat that makes the higher alacrity later feel like it has even more of an effect on making things easier.

[XtremJedi]

Aye indeed.

 

And also that there are some situations where being able to add in extra abilities between heals will increase your ability to maintain easy healing on the tank. For example, being able to Triage effects or use Flash Grenade to lower incoming damage on the tank from a group of mobs, while DPS burn off extras. Or being able to interrupt abilities or being able to re-slice a droid in combat etc... or even, having more time to throw in some DPS on trash when you're healing quite comfortably.

 

Not everything is immune and it's not only about boss fights... though I recognise that this is partially a play-style choice. Some healers see Interrupts and CC as being a DPS role and think that healers need to focus entirely on healing.

 

Some people do seem to see you as the fount of all knowledge around here though... so it might be worth including something in your OP that acknowledges these other, more unpredictable (or even incalculable?) benefits. I know some of it may seem implicitly true - but there are many people who will read your post, who perhaps won't reach those conclusions without a pointer.

 

X

[RuQu]

Aye indeed.

 

And also that there are some situations where being able to add in extra abilities between heals will increase your ability to maintain easy healing on the tank. For example, being able to Triage effects or use Flash Grenade to lower incoming damage on the tank from a group of mobs, while DPS burn off extras. Or being able to interrupt abilities or being able to re-slice a droid in combat etc... or even, having more time to throw in some DPS on trash when you're healing quite comfortably.

 

Not everything is immune and it's not only about boss fights... though I recognise that this is partially a play-style choice. Some healers see Interrupts and CC as being a DPS role and think that healers need to focus entirely on healing.

 

Some people do seem to see you as the fount of all knowledge around here though... so it might be worth including something in your OP that acknowledges these other, more unpredictable (or even incalculable?) benefits. I know some of it may seem implicitly true - but there are many people who will read your post, who perhaps won't reach those conclusions without a pointer.

 

X

 

Some things, like reapplying CC, aren't really model-able.

 

One thing that is worth pointing out, and that I think many people forget/miss/ignore is that larger heals also increase your downtime. It's important to try and get a sense of how much of an improvement we get from each and not falsely assume that only the ability that explicitly reduces cast time increases our time not casting.

[LancelotOC]

Ruqu I have to ask, since you are mentioning it again and again, what do you mean by increases our downtime. Are you only speaking for Commado heals?

 

Or are you refering to the sage big heal and saying it creates inherant downtime because of the power it costs or are you saying the sage big heal with alacrity is less efficient because it costs more per second since it now casting faster?

 

As the later would seem mute to me since the force costs is the same and casting it any faster would simply mean you would start regenerating the mana cost earlier. The end effect must be better unless alacrity somehow effects our regen rate negatively?

 

For the former are you saying a big heal is less efficient than other heals? Which heal are you talking about. Sorry if you clarified this earlier as I haven't read the whole thread.

 

Love your graphs.

[RuQu]

Ruqu I have to ask, since you are mentioning it again and again, what do you mean by increases our downtime. Are you only speaking for Commado heals?

 

Or are you refering to the sage big heal and saying it creates inherant downtime because of the power it costs or are you saying the sage big heal with alacrity is less efficient because it costs more per second since it now casting faster?

 

As the later would seem mute to me since the force costs is the same and casting it any faster would simply mean you would start regenerating the mana cost earlier. The end effect must be better unless alacrity somehow effects our regen rate negatively?

 

For the former are you saying a big heal is less efficient than other heals? Which heal are you talking about. Sorry if you clarified this earlier as I haven't read the whole thread.

 

Love your graphs.

 

Sorry, I am referring to downtime as a positive thing, which is bad wording on my part. I should probably have said increased "free time" , "relaxation time" , or "opportunity time."

 

If you cast a big heal faster, you can then relax afterwards.

 

If you make that big heal bigger, you can instead wait longer before casting it.

 

Both allow you extra time to make decisions, cast utility, etc. It is just a question of return on investment to see which buys you more.

[XtremJedi]

Some things, like reapplying CC, aren't really model-able.

 

Hehe... I have a bit of deja vu.

 

I suppose the question is: what is your goal with this analysis?

 

If it is your goal to provide people with the necessary information to choose whether to put Surge or Alacrity in their kit (as an example) - then it seems to me, that you can't really answer this question if you're leaving out a range of factors, because they're not model-able or the situations and benefits haven't been defined well enough.

 

If it is your goal to simply provide a restricted analysis of which stat is better for pure healing throughput, given a restricted set of assumptions... then that's all well and good, but it doesn't really answer the question of which stats are better.... and I think that many people reading your OP in this thread, might well assume that this is meant as a complete analysis which will tell them which stats are better to equip.

 

Perhaps you might add a line to the conclusion along the lines that the information provides useful data that can go towards deciding what to equip - but that there are several factors that remain outside the modelling and so the results cannot be seen as a definitive answer to which stat to equip? Or you might include a note of some of the other potential benefits that aren't modelled for, and encourage people to include these in their own evaluations.

 

I'm not married to any of these stats at all. It's not that I feel that alacrity is particularly better. I personally don't go out of my way to stack alacrity, though I recognise some of the benfits and do use an alacrity relic. I'm really just pointing out potential benefits that seem to have been missed and asking whether it might not be worth making the frame of reference for your data a little clearer for those who might not read with such care and deep understanding as you.

 

X

Edited March 12, 2012 by XtremJedi

[RuQu]

I guess part of what I was trying to say there, by clarifying the point about power/crit/surge also increasing healer free time, was that I don't think I really need to model those utility things like CC, etc, with this metric.

 

If it was a HPS metric, then it wouldn't really say much. But HPRCT says that as your HPRCT increases, you will gain more time for whatever utility you want. For some people, that means more time to serve as RL and call out things. For others, it means tossing in some DPS. For yet others it means tossing out interrupts, stuns, CC, etc.

 

What you do with your free time is up to you. That's why its free time

 

*edit*and if you want to blow all of your Ammo partying with Charged Bolts during your free time... well that's not really any of my business, is it?

Edited March 12, 2012 by RuQu

[Obtena]

I want to thank RuQu for the work and the thread. I always suspected that Alacrity was not being given it's due respect. One thing that I think would benefit the analysis is some consideration of the impacts of being interrupted, specifically in PVP. It's my suspicion that as you increase the downtime of an operative, alacrity benefits doubly in ways that power/surge and crit can't impact:

 

1. Harder to interrupt faster executing actions

2. Faster recovery after interruptions

 

I have to admit I didn't follow your work with rigour but I didn't see an attempt to factor that in as as a PVP healer, that's the kind of information I would like to see. You can even imagine that it impacts PVE due to interuptions that people don't typically think about (moving around for position, etc...)

 

P.S. I see you say that modeling interuptions isn't possible but I think doing some simplified attempt like adding a few seconds to a small percentage of heals might be a good simulation.

Edited April 3, 2012 by Obtena

[RuQu]

I want to thank RuQu for the work and the thread. I always suspected that Alacrity was not being given it's due respect. One thing that I think would benefit the analysis is some consideration of the impacts of being interrupted, specifically in PVP. It's my suspicion that as you increase the downtime of an operative, alacrity benefits doubly in ways that power/surge and crit can't impact:

 

1. Harder to interrupt faster executing actions

2. Faster recovery after being interrupted

 

I have to admit I didn't follow your work with rigour but I didn't see an attempt to factor that in as as a PVP healer, that's the kind of information I would like to see.

 

The big reason is that I don't PvP. I haven't done a single WZ, so I don't have the requisite knowledge to even comment meaningfully.

 

The second is that PvP is extremely difficult to model. I haven't even tried.

 

As a non-PvPer, I'm not even sure how PvP interrupts work. I know in WoW it locked out a whole school, so if you could interrupt a shaman's heal you prevented any heal since they shared the same school. With SWTOR, if I interrupt Adv. Med Probe, can you still cast Med Probe or Bacta Infusion? How long is the lockout? Just a GCD? Is that affected by the cast length of the ability? If the ability has a CD, and gets interrupted, does the CD apply?

 

At the very least, I would think that being faster would make it harder to interrupt against a bad to mediocre PvPer. A good one can still reliably interrupt even a 1s cast time ability, so it wouldn't help much there. On the other hand, some interrupts, like displacement abilities, have long animations (like the Shadow/Sage knockback) and a faster cast might finish before the animation interrupts you. Certainly worthwhile from that point of view, and you will spend so much time stunned/running/interrupted that you can probably regen the loss in EC.

[Kallti]

Whenever I see Xtreme and RuQu responding to each other in a thread, I imagine two old codgers sitting in a wood-fire heated lounge room somewhere in country England wearing smoking jackets and puffing on pipes while chuckling to each other; "I say, old chap, jolly good graph you have there". "Ho, oh, thanks old chum".