Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

[Theorycrafting] Sage Heal Stats

STAR WARS: The Old Republic > English > Classes > Sage / Sorcerer
[Theorycrafting] Sage Heal Stats

Nibbon's Avatar


Nibbon
06.07.2013 , 02:52 PM | #21
Quote: Originally Posted by Darth_Dreselus View Post
A 16 man DG with 67% overheal. Well it is a 16 man and Sages will always have fairly high overheal from Salvation, but what that log shows is that spam-Salvation is possible, even viable but hardly ideal.

Only 16.36% of heals were directed at tanks, including one Force-bent Dark Heal with 100% overheal each. My guess here is either fat finger, double right click, mistake or the tank took a big hit but the heal was sniped by another healer before the Dark Heal landed. The use of the other 6 Dark Heals seems correct and likely in that fight. Again this tells me he was predominantly a group healer, which explains the off cd Revi. 14 Innervates were cast at one of the tanks though so he may have been filling a double role. (AoE DPS every 15s, Focus on a tank) It is hard to tell from the log.

Most Revis are not Force-bent, in fact only one is. In line with general Sorc healing practice and at least it dipells my fears of there being some hidden rotation I can't figure out. Due to hitting 3+ people cost is more than justified (provided it was Effective can't really tell).

It's just a strange log to use as a base of Salvation frequency.

I sadly do not have a 2.0 log of DG. In fact the only one I have is from my first clear of TFB so the timing of heals is different from the one I use now when I predict big hits and AoE damage due to experience.

http://www.torparse.com/a/226118

The usage of Salvation and EHPS ratio speaks for itself. OC is useless because of the nature of the fight (and a CCed droid) and Olok is a weird one I do think I just spammed Salvation. I do know I DPSed P3. Trasher, Warlords and Styrak are pretty good fights as far as 'general healing' goes (AoE DPS, focus tank).

So I played with your spreadsheet using different set ups (Salv every 30s, NS every 7.5s, no Salv, no Salv+no FA) and I always get 5/5 as best HPS. Very interesting, will have to play with the other spreadsheet to see where the discrepancy is.
Again, I don't feel like arguing, so excuse my lack of a response here.

It wasn't a basis for salvation frequency as much as it was proof that it could be cast that much. He was definitely a raid healer in this fight - and it is one of the few fights I know of that requires constant healing (well when it came out, its very easy now). I agree most aren't force bent, but can be, I just think he didn't need to for a HM. I agree Thrasher might be a better fight to use, but its a very very short fight. DC, in this log, was the exact 5 minute length I was hoping for.

All I know for sure is Ace is an excellent guild, so if they like what I'm doing I will be happy with myself.

I'm glad you got to play around with it - I think 5/5 is a good mix, not so different from my 6/4 surge/alac, anyway. 5/5 is what I originally came up with.

I can explain the discrepancy, but I haven't taken a good look at the other spreadsheet. The obvious answer is that alacrity is being way overvalued. I have some graphics I have been preparing in order to try to explain it, they are pretty much ready now ... see below.

______________________________

Alacrity is effected by a number of things - it's own stat formula - then the cast time formula - then a lesser known issue that I have been exploring, the effect of cooldowns on alacrity.

Diving straight into the numbers, let's look at the individual increase of HPS for one alacrity enhancement at the underworld level.
79 alacrity=1.15% alacrity percent - will use the GCD, but any placeholder number works, like 1, or any other cast time.
1.5/(1+.0115)=~1.4543
1-(1.4543/1.5)=1.13% increase in cast speed (or cast time reduction)
1.13% is not the increase in HPS
cooldowns aren't effected by alacrity, except in that the spell being casted when the cooldown is off casts faster. Let's just call this the alacrity cooldown coefficient. The coefficient is a dynamic number that goes up (reducing the severity) as you add more alacrity. With 79 alacrity, the coefficient is roughly .6185.

To find the effect on actual HPS from reduced cast time, you multiply the reduced cast time by the coefficient
1.13%*.6185=~.7%

Surge also has a coefficent (this is well known). This coefficient is static, not dynamic (unless you alter crit in different ways). The Surge coefficient= (# spells effected by crit that are casted)/(Total # of spells casted) * critical percentage

I'll go into the results again, but I wanted to show, graphically, what the alacrity cooldown coefficient does. The following image is a very specific case, 2 spells being casted, one with a cooldown, both being effected by alacrity. The graph shows the relationship between the points of alacrity, the reduced cast time, and the effect that the reduced cas time mixed with the effect of the cooldowns has on actually increasing your HPS. Again, this is a very specific case which I made purely in order to demonstrate how cooldowns CAN effect HPS (and alacrity specifically):
https://docs.google.com/spreadsheet/...x=qobq6k5bo1dc

This specific case can be found here: https://docs.google.com/spreadsheet/...xWWNBZ2c#gid=2

I do not recommend looking through it - I didn't exactly design it to be friendly to read or alter.

So, going back to the real spreadsheet I worked out (https://docs.google.com/spreadsheet/...xWWNBZ2c#gid=3), you can look at the two stats individually. If you had 0 enhancements of each and increased one or the other by 1 at a time, you would slowly be increasing your HPS (that is a given). So this is the effect if you were slowly adding just surge or alacrity points and how much each would increase your HPS:
https://docs.google.com/spreadsheet/...x=rnnlqwmpnsvv

Realistically, you will always have 10 "enancements" or in the underworld level, 790 of the total secondary stat. This is how the two interact with each other - the height of the line is the 6/4 surge/alacrity number I have been basing BiS off of.
https://docs.google.com/spreadsheet/...x=2g8byv8x8ekq

Now, given, if you change my model and frequency of cast of some of the spells - the numbers will change. If you threw in crit (Which you shouldn't) the numbers will change too (crit = higher surge coefficient). That being said, you can pretty much drastically alter it and will still get roughly the same number (5/5, being the one exception). It is important to understand why alacrity isn't as good as it looks before comparing to surge (which is a lot more straightforward).

Darth_Dreselus's Avatar


Darth_Dreselus
06.07.2013 , 03:24 PM | #22
Quick question the orange line is the instant heal?

And yes this is purely 'academic'. I'm pretty sure even a 10/0 setup is enough for clears. That does not mean that I still don't want to have min/max gear, if nothing to feel good about it.
No One Important Died Tavernus - Sorc / Taverna - Mara / Pintus - Op / Gintonica - PT
The Red Eclipse. Sweet Dreams to all my Harbinger children.

Nibbon's Avatar


Nibbon
06.07.2013 , 04:05 PM | #23
Quote: Originally Posted by Darth_Dreselus View Post
Quick question the orange line is the instant heal?

And yes this is purely 'academic'. I'm pretty sure even a 10/0 setup is enough for clears. That does not mean that I still don't want to have min/max gear, if nothing to feel good about it.
The orange line in the first link? That is the alacrity cooldown coefficient that most of that bottom half of my previous post is about. The orange line is the actual benefit to your HPS from the percentage of alacrity along the same x-axis.

And while that second paragraph was for someone else, I agree with / understand the sentiment

Darth_Dreselus's Avatar


Darth_Dreselus
06.07.2013 , 04:11 PM | #24
Quote: Originally Posted by Nibbon View Post
The orange line in the first link? That is the alacrity cooldown coefficient that most of that bottom half of my previous post is about. The orange line is the actual benefit to your HPS from the percentage of alacrity along the same x-axis.

And while that second paragraph was for someone else, I agree with / understand the sentiment
Oh I see the two casts in the sample are combined.

Well, it may have took a page of back and forth and while I would not say that I am convinced I would definitely want to see the hypothesis examined by either Noim or Orderken. Figuring out whether it is mathematically sound, let alone correct is beyond me. For instance I have no idea where you got the cooldown coefficient from (even after scanning formulas and delving into the first page of the spreadsheet).

If it proves to be correct, bravo to you Nibbon. Will it also affect the DPS gearing or is the fact that half the slots are taken up by accuracy too restrictive?
No One Important Died Tavernus - Sorc / Taverna - Mara / Pintus - Op / Gintonica - PT
The Red Eclipse. Sweet Dreams to all my Harbinger children.

Aurojiin's Avatar


Aurojiin
06.07.2013 , 04:24 PM | #25
Don't get me wrong, I'm not implying that chasing BiS gear is worthless, or disparaging anyone for it (after all, I do). It's just that, sadly, I lack the motivation to contribute any more.

That being said, I have to question your hypothesis, NIbbon. In my experience, typically one already has to choose between idling and delaying a higher priority ability past CD because of a filler delaying it. You seem to be saying that higher alacrity results in a statistically significant increase in the number of times when you would delay something like HT or Salvation. I haven't found this to be the case in practice.
Aisev -:- Seer Sage Si'ki -:- Darkness Assassin

Nibbon's Avatar


Nibbon
06.07.2013 , 04:29 PM | #26
Quote: Originally Posted by Darth_Dreselus View Post
Oh I see the two casts in the sample are combined.

Well, it may have took a page of back and forth and while I would not say that I am convinced I would definitely want to see the hypothesis examined by either Noim or Orderken. Figuring out whether it is mathematically sound, let alone correct is beyond me. For instance I have no idea where you got the cooldown coefficient from (even after scanning formulas and delving into the first page of the spreadsheet).

If it proves to be correct, bravo to you Nibbon. Will it also affect the DPS gearing or is the fact that half the slots are taken up by accuracy too restrictive?
The coefficient isn't exactly a real thing, per se. It was just my best way of explaining it. What is really happening is the cooldown timing, it creates the "dynamic coefficient". Again, the most basic explanation is - your best abilities have a cooldown, so you have to use a filler between cooldowns, thus your HPS (or DPS) does not increase by the cast time - it still has to account for the cooldowns and when those spells are specifically casted (and how much the lesser heals actually hit for in comparison to the bigger heals). Obviously saying that there is a dynamic coefficient is easier, even if it is a half truth. The reason it is true is because there is always a number you can apply to another number to reach ... any number. The mere fact that it is dynamic makes it not exactly a coefficient. Sorry if I am being extremely confusing, but I assure you that I have a full understanding of it in my own head even if I have trouble expressing it to others.

As for DPS - it is already taken into account: http://www.swtor.com/community/showthread.php?t=627405 That is my thread that I created at the beginning of 2.0. No one has really disputed it as far as I am aware. Obviously DPS is a lot more simple and straightforward than healing because you actually have a real rotation to follow. The only trouble I have proving that it exists for heals too is because healing is a much more dynamic activity DPS also has the advantage of dummy parses - of which I do quite well on myself, but also the people following my advice tend to be up towards the top. Obviously there are exceptions, but beyond just blaming them on luck, I can mathematically prove they are luck by drawing a line between expected crit damage and actual crit damage. I can go on about that for a while, but hopefully that is sufficient enough of an explanation as needed ;p



edit: p.s. - I don't know Noim or Orderken, but happy to discuss my analysis and methodology with them.

Nibbon's Avatar


Nibbon
06.07.2013 , 04:36 PM | #27
Quote: Originally Posted by Aurojiin View Post
That being said, I have to question your hypothesis, NIbbon. In my experience, typically one already has to choose between idling and delaying a higher priority ability past CD because of a filler delaying it. You seem to be saying that higher alacrity results in a statistically significant increase in the number of times when you would delay something like HT or Salvation. I haven't found this to be the case in practice.
Ew, I don't like idling. There is never a break in casting in that analysis - what you have is an overlap. The overlap is the cast time going off from the non-cooldown spell - the spell coming off cooldown and then being casted - resetting the cooldown timer. I did a separate analysis that even a very slight idle is worse for HPS than filling with a lesser heal. Call me lazy, but I've already done a lot of work on this so I am not going to recreate it

I disagree entirely with your statement, though. I am not suggesting anything gets delayed more often. I admit that in the hypothesis page it would look like that, and may be overstating slightly the effect (but the effect exists regardless). On the HPS spreadsheet, though, we have partial heals (as in, 55.15 rejuvenations, for examples). Those partial heals are basically the lack of an overlap - so it really diminishes the "delay" you are worried about. So the fact that alacrity takes such a huge hit seems all the more real to me.

One note about things in practice. Besides ability lag - it is difficult to see how much of your spells are overlapping. The way the UI currently works is - if a spell will come off CD when another spell is going through the GCD cycle - it will appear like it is ready - in reality you can't interrupt an ability that casts over the length of a GCD (I mean you can, but you can't begin casting something in that time, you have to wait for GCD to finish). What I'm saying is - it might look like you aren't overlapping, but you definitely are, even if by fractions of a second (which is really what we are talking about here).

Darth_Dreselus's Avatar


Darth_Dreselus
06.07.2013 , 04:46 PM | #28
Quote: Originally Posted by Nibbon View Post
I assure you that I have a full understanding of it in my own head even if I have trouble expressing it to others.
Well that's...reassuring I guess

Quote: Originally Posted by Nibbon View Post
DPS also has the advantage of dummy parses
This. So much of this. I have actually tried to parse at least ST healing.

Best way I came up was grouping with a tank and have him constantly taunt some low level mob and I just heal him with an efficient rotation, then like a 10s burst, 30s burst etc. Works pretty well on Scoundrels and Commandos. But I wanted to know about Sages. Now I am 'wasting' a GCD on NS but that's fine Scoundrels have to DS and Mandos have to HS. The problem is that it takes away my health so my limitation was always going to be HP, Force or the combination of the two. And If I skipped the NS it would not be an efficient rotation so all I could measure was burst. So instead I just idled for a GCD instead but I still could not get the 5 min parse I wanted. So I then used Salvation which I guess kind of worked but it again was not the rotation I wanted.

I later realised I could just have another healer topping me up (possibly even a companion would work). But I just waved my hand and went with the old "as long as content is cleared that's all that matters".

This was pre 2.0 I guess I could try again with the heal from Force Armour.
No One Important Died Tavernus - Sorc / Taverna - Mara / Pintus - Op / Gintonica - PT
The Red Eclipse. Sweet Dreams to all my Harbinger children.

Aurojiin's Avatar


Aurojiin
06.07.2013 , 04:46 PM | #29
Perhaps I'm fundamentally misunderstanding your hypothesis. Are you saying that Alacrity's efficacy with respect to HPS is reduced because you have to use a larger number of inferior filler abilities while waiting for cooldowns?
Aisev -:- Seer Sage Si'ki -:- Darkness Assassin

Darth_Dreselus's Avatar


Darth_Dreselus
06.07.2013 , 05:00 PM | #30
Also I think I am going to load up TOR, get an EV going and just hit some keys on my keyboard and watch the pretty green numbers after all this.
No One Important Died Tavernus - Sorc / Taverna - Mara / Pintus - Op / Gintonica - PT
The Red Eclipse. Sweet Dreams to all my Harbinger children.