[Resto] spreadsheets and calculations - Page 3

Nagisamuro · 03/05/09, 10:52 AM

Originally Posted by Daidalos

I've updated my sheet a bit. I've added in HPS HEP calculations for lhw, hw, and ch. I left out riptide due to the hot nature of it and it being a small part of our healing but if people really want to see it I can include HEP for it as well.

I am debating adding in HEP for regen but it really boils down to crit and replenishment and mp5 as our only regen so its going to vary on your raid make up and rotations a lot. Do people want this or is this more of a based on my experience kinda thing?

Also I added in the glyph for earth shield assuming it a 1.2 multiplier after all other calculations are applied.

I appreciate all the effort you've put in to this spreadsheet, Daidalos. It has been of both practical and academic value as I've geared up my shaman and tried to understand the class mechanics better over the past several months.

Regarding the HEP calculations: how do you suggest one determine the percentages to use for the three spells (Overview!B20:D21), and the weights for HEP & MEP values (I20, I21)? I think I understand what the value are for - weighting the stats that are most beneficial to the spells you use the most, and then weighting the importance of throughput vs. longevity - right? The spell weights I have been filling in based on my own mental aggregation of the breakdowns reported by Recount for my healing performance in raids over the past few months, but the values for HEP/MEP weighting ...I'm having trouble figuring out what to base those numbers on, and they can have a significant impact on the final HEP values.

Daidalos · 03/05/09, 1:17 PM

Originally Posted by Nagisamuro

I appreciate all the effort you've put in to this spreadsheet, Daidalos. It has been of both practical and academic value as I've geared up my shaman and tried to understand the class mechanics better over the past several months.

Regarding the HEP calculations: how do you suggest one determine the percentages to use for the three spells (Overview!B20

21), and the weights for HEP & MEP values (I20, I21)? I think I understand what the value are for - weighting the stats that are most beneficial to the spells you use the most, and then weighting the importance of throughput vs. longevity - right? The spell weights I have been filling in based on my own mental aggregation of the breakdowns reported by Recount for my healing performance in raids over the past few months, but the values for HEP/MEP weighting ...I'm having trouble figuring out what to base those numbers on, and they can have a significant impact on the final HEP values.

To me I don't really like the evenly mixed weights. However, If you just want one set then I would go off of whats the ratio of healing spell power : mp5 for you. I don't really know how to make it more intuitive due to healing's very situational nature.

What I've been doing is make one heavily weighted towards HPS and then look at the loot then make another heavily regen weighted and look at the loot. Since we have differing needs and some are even conflicting (e.g. haste vs regen) I personally would go with different sets. Also keep in mind that your raid buffs will affect the weighting as well so not only do you need to think about what you do on different fights but what buffs other people give you. I am also finding that sometimes I want both regen and haste for bursty fights that require a lot of healing then have time to regen before the next burst. SO I created the ability to limit haste's negative MEP values.

Handyhoof · 03/06/09, 9:22 AM

I have been recommending "gut feeling" for throughput/longevity coefficients. I really feel that you can't get that number from someone else. On my sheet they default to 1 and 1, but are meant to be weighted based on personal preference/style/deficiency.

Daidalos · 03/06/09, 11:52 AM

Originally Posted by Handyhoof

I have been recommending "gut feeling" for throughput/longevity coefficients. I really feel that you can't get that number from someone else. On my sheet they default to 1 and 1, but are meant to be weighted based on personal preference/style/deficiency.

I would like to think its a little more analytical than a "gut feeling" but I do think your weights will depend on what situations you tend to be in rather than what others do. For example I tend to find myself the CH bot still, only interrupted to renew my ES on a tank, (we rarely have a healing priest or druid but we have several good pallys) so I tend to look for spell power, int, haste, and mp5. Since we don't run heavy on replenishments (1-2 in 25 man) I can have mana problems if everything goes south but under normal conditional with the 2pct6 bonus I can spam ch the entire fight. So with spellpower and haste I am looking at good HPS gains and I balance things a little with some int and mp5 for those "everything that can go wrong will go wrong" situations.

Siggidzweie · 03/20/09, 11:51 AM

I already posted that in another topic, but it should be here too.

The upcoming wall-of-text is a little bit off-topic and belongs actual to another topic, but I'll post it here too.I already asked that myself months ago, but decided to hold back, cause I didn't wanted to mess with Daida and Skyhoof, who really do a great job for our resto-community (props again!).... but still .... criticism can't be always positive and the more time passes by, the more comments concerning some numbers of healing-spreadsheets and -rotations I see and that's, although it's the most simple way ("don't think about it, just put in your numbers and become a superstar"), simultaneously the wrong way to decide how to gear up or to play.
I really respect Daidalos time, work and knowledge, that he invested in this formular, but it's use seems to be really, really superficial and dubious. It's not Daidas work or skill, that makes it become superficial I really admire his accuracy and knowledge, it's the nature of healing. Spreadsheets and sims make sense for Damagedealer, because Damage in most kind of situations isn't capped, exept by the lack of your own skill or movement-abilities, so you could basically rely on those numbers and take them as a kind of benchmark.
Healing-Output depends on so many more variables that are capricious. I'm not only talking about evident tasks like "Do I need to stack int, cause it gives the greatest benifit for a shaman, although I end up with 10k Mana on every bossfight". I'm talking about design of encounters, like is there much raid damage, do raidmembers need to spread out, a.s.o and other things like if you have skilled healing combatants in your crew, who know their job too, witch will. definitly reduce your own amount of healing and finnaly the unpredictability of human actions.
Also static rotations are in most kind of situations (except single-target-healing-situations in 10-mans and in situations, where CH ist mostly useless) useless for us as a raid-healing-class. I know, that tapping out the full potential of spell-synergies (RT <-> CH; RT <-> HW/LHW, aso.) is a seductive idea. I myself thought about that, when the new spells were released 9 months (or what ..) ago, but I had to adjust myself later on, on the same reasons I regard spreadsheets for unhelpful. You should rely more and repetitive on your own infight-expiriences, keep your eyes wide open, on what you've done in different situations and different setups and compare it with your recount/wws-stats to check out the result. And ... a good healer needs to know his class and his job in different fights. That's why the resto-wiki is a gorgeous intruduction for newcomer. Every healer who wants to become top-notch has to drill himself, cause that's the true fundament. That's a good way to find and fill out your own "healing-niche". A spreadsheet in most cases can't do that for you.

Daidalos · 03/20/09, 12:39 PM

Originally Posted by Siggidzweie

I already posted that in another topic, but it should be here too.

First off, ouch my eyes. Could you format that so its human readable or something?
Secondly spreadsheets aren't to tell you how to play. However once you know how you play you can then check the spreadsheet for math about specific questions. For example "would 15 crit rating be more hps than 7 spellpower?", "would losing 4pc t7 bonus be worth gaining 100 spell power?" or "is 50 crit rating or 10 haste better for Chain heal?".

I have on numerous occasions been the proponent of pay attention, and do post fight analysis with wws or other parser tools. I've also said that strats and people paying attention make a much bigger impact on the raid than your itemization. As such I'm not sure who your post is aimed at. I think everyone in here already knows these points. If people think you should always do whatever is the highest hps rotation regardless of the situation then the person has completely missed the point of spreadsheets.

Spreadsheets answer specific questions about your situation, not how to play, or where do I stand. I think this goes without saying.

mikeymoollet · 03/22/09, 8:47 PM

in reference to crit regen, it is too unreliable of a source of mana to take into account due to it not proccing off of every heal. the most mana regen i have seen from imp water shield(total water shield mana gain-water shield base mp5/(time in seconds*5)) is roughly 370 mp5 which is about the same as 100% replenishment uptime. i was geared for crit to see the maximum potential of water shield at the time so this is unreliable. crit regen is a very good source of regen to help meet the increased mana usuage from hw/lhw spam. making crit a very valuable resource.

Vuldunobetra · 03/24/09, 1:33 PM

Assuming Purification, Tidal Waves and Improved Chain Heal, these are the formulas I got:

LHW = 1892 + 1.00 * SP [corrected]
HW = 3575 + 1.99 * SP
CH = 1492 + 1.77 * SP (1 target)
RTd = 1837 + 0.43 * SP (direct heal)
RTh = 1835 + 1.05 * SP (HoT total)
RT = 3674 + 1.47 * SP

Chart at http://spreadsheets.google.com/pub?k...3&output=image [reposted with corrected values]

This confirmed my gut feeling that the initial heal of Riptide is not as good as it seems when I dinged 80. It is not scaling nearly as well as our other spells.

Daidalos · 03/24/09, 2:15 PM

Originally Posted by Vuldunobetra

Assuming Purification, Tidal Waves and Improved Chain Heal, these are the formulas I got:

LHW = 1892 + 1.88 * SP
HW = 3575 + 1.99 * SP
CH = 1492 + 1.77 * SP (1 target)
RTd = 1837 + 0.43 * SP (direct heal)
RTh = 1835 + 1.05 * SP (HoT total)
RT = 3674 + 1.47 * SP

Chart at http://spreadsheets.google.com/pub?k...3&output=image

This confirmed my gut feeling that the initial heal of Riptide is not as good as it seems when I dinged 80. It is not scaling nearly as well as our other spells.

I take that you aren't including AA, but even so my formula aren't matching yours.

for example
LHW:
[base +SP*[ (1.88)*(1.5/3.5)+.1] ] *1.1
or base *1.1 + .SP*.996

Edit: Could you repost that graph? I think the old incorrect one might confuse people.

Vuldunobetra · 03/24/09, 2:21 PM

Yes, I had a formula wrong in LHW, now corrected. I spot checked the others against your SS. Looks right, off by a point or two due to rounding.

[later]

For HW+HW, I get: [ base + SP * (1.88 * 3.0/3.5 +0.2)] *1.1 *1.18
HW+HW = 4218 + SP * 2.351

Holy Light = [ 5166 + SP * 1.66 ] * 1.12
Holy Light = 5786 + SP * 1.859

Greater Heal =(base + SP * (1.88 * 3 / 3.5 + 0.4))*1.1
Greater Heal = 4697 + SP * 2.21

All three spells are almost equal at 3000 SP (11,263 11,395 11,327) I also haven't counted AA in these numbers, nor took into account mana costs, cast times, crit, and other talents. And assuming I haven't botched another formula.

stassart · 03/30/09, 12:52 PM

Hi Daidalos,

I've been going over your spreadsheet and comparing your equations against my own for my next release (my next release will hopefully provide HEP values for talents and glyphs so you know what you are getting from the talents and glyphs you have picked). My current release doesn't ask about all talents, so this should improve the quality of the results a bit (so far the HEP results are only changing out 2-3 decimal places for me, but the more accurate the better).

Except for the value of MP5, I am changing the way I find the most mana intense fight as my old method was good at finding long fights that used a lot of mana, but poor for short mana intense fights like patchwork. I have done a bit of reitemizing and regeming based on my results from shaman_hep and had mana problems on Patchwork (spamming riptide/healing wave/healing wave) and have to use more pieces from my mana regen set for that fight, so I fixed my algorithm based on that experience contradicting the data from shaman_hep.

A couple of questions I have from your equations,

For riptide, you use the equation:

(1+Stats1.$E$19) *(((($B$3/3.5)*0.5)*Stats1.$E$38*1.88)+D3)

Which unless I am reading it wrong, uses the value from $E$38 for spellpower, this field defined is:

Total +healing to Lesser Helaing Wave with relic

I would have thought riptide would use:

Amount of spellpower + SP buff

which is $C$7.

Is this wrong, or does Riptide actually use the lhw relic slot to effect its value? It does not seem to based on a quick test in game, but that isn't a lot of data to work with.

My other question is based on the Earth Shield equation:

(D3+(Stats1.C7*1.88*0.2865))*(1+Stats1.E19)*(1+Stats1.E25+Stats1.E33+0.2)

Which to me reads as:

Amount per charge (base spell)

+ ((Amount of Spell Power raid buffed) * 1.88*0.2865)
(This works out to a coeffecient for 0.53862, I use 0.5377 according to wowwiki, both are close enough that it is hard to tell which is correct)

* (1 + Purification %)

* (1 + Improved Shields + Improved Earth Shield + 0.2)

What is the 0.2 from?
Where is this extra 20% coming from?

It can't be the Glyph of Earth Shield in 3.1, because you have a separate entry for:

w. crit and glyph

which uses:

D7*(1+Stats1.$E$62)

$E$62 is the extra 20% from Glyph of Earth Shield.

Thank you very much for your work on your Spreadsheet, it is nice to have something to check against.

Daidalos · 03/30/09, 1:18 PM

Originally Posted by stassart

Hi Daidalos,
For riptide, you use the equation:...
My other question is based on the Earth Shield equation:...

Looks likes those are bugs I"ll check.

Edit: Thanks for spotting those. I confirmed they were bugs and should be fixed now. When you see things hard-coded like they were in the earth shield calcs it usually means I was messing with something then forgot to take it out (I guess I should refrain from auto publishing but I figure its better to have instant updates and have other eyes finding bugs than just mine). When the patch notes about the new ES glyph came out I was playing with numbers to see the results and totally forgot about it. Riptide was most likely a copy n paste bug that I never noticed since I never use lhw relic.

stassart · 03/30/09, 5:30 PM

Hi Daidalos,

Sorry to bug you again, but your equation for healing stream totem does not seem to match the in game values. You actually have a value that is almost double what I see in game.

You have:

(C2+(Stats1.C7*1.88*0.035))*(1+Stats1.E19)*(1+Stats1.E18)*(1+Stats1.E58)

First of all you use a coefficient of 1.88 * 0.035. Mine was wrong too, but I think this yours may just be a typo, using my data I work out to a coefficient of 0.846. That is 1.88 * 0.045. The old pre-spellpower change coeffecient was 0.045 (which is what I have in the current release of shaman_hep cause that is what wowwiki lists with a warning that it might be wrong) so that makes sense.

Which if I read it correctly works out to:

(base amount (25)
+ (
(Spellpower * coeffecient)
* (Restorative Totems)
* (Glyph of Healing Stream Totem)
)

To match my in game data, it should be:

((base amount
* (Restorative Totems))
+ (
(Spellpower * coeffecient)
* (Glyph of Healing Stream Totem)
)

((C2*(1+Stats1.E18))+(Stats1.C7*1.88*0.045))*(1+Stats1.E19)*(1+Stats1.E58)

I have not tested with Glyph of Healing Stream totem myself.

The effect of Restorative Totems seems to be on the base spell with no effect from spell power. This may be a bug, but it definitely makes the talent completely gimp in its effect on healing stream totem.

Daidalos · 03/30/09, 6:44 PM

Devs have acknowledged they are looking into the restorative totem bug. I am waiting for this to be fixed or for blizzard to state that it is working and intended. I'll fix any inaccuracies at that time.

Edit: I've decided to put the coef back at .44 I had it at .35 when I was trying to figure out why the initial numbers from restorative totems was so low. ( I was assuming the talent still worked so I ended up with a lower coef) for now I"ll leave the bonus working as I htink it should with the proper coef so people can TC what HST will do.

stassart · 04/04/09, 9:52 PM

Hi Daidalos,

For the effect of riptide on chain heal you have:

F8*1.2

Shouldn't this be F8*1.25?