[Paladin] Holy raid itemization for best performance

While this may be possible to improve, currently there is no better known way to evalulate stats. As in, this method takes into account more than anything else I've seen so far.

If you use up the buffer of 300 mana and it can be refilled, you can use it again, effectively doubling the use you get out of your buffer and pushing it higher than simple mp5. Is there something wrong with this idea?

If you ever get your mana back to full by some wierd fight mechanic (which really doesn't happen on any fight I can think of except RoS where your mana hits 0 anyway), in a fight where you're possible to go oom before that, I'd still just count it as 2 seperate fights, and calculate the value of my stats for fight #1 and then the value of my stats for fight #2, and weight them based on what stats would help me more A. on average B. on the fight that actually matters more.

You'r talking about a phased fight of some kind where you get full mana in between bursts of healing? In the case of RoS I agree that int has much more value then in any other fight.

In any other normal fight where you spend mana in a rather constant flow and regain it in a rather constant flow, that 300 extra mana is doing nothing to your regen unless you hit a point mid fight where your mana pool is completely filled again while actually having spent the 300 mana, making sure you have less 'wasted' regen.

There are maybe 2 fights that I can think of that come close to those mechanics aswell, being Kael and Illidan (both times mostley phase 1). But due to the nature of the other phases of those fights, I wouldnt value int more then in any other fight.

If you actually reach full mana mid-fight without some special mechanic to it, it means you didn't really need to do much healing for a while, which again can allow you to just seperate the fight into phases and check how good your gear is for each phase seperately. Having higher starting mana doesn't really change how effective regen is, though, as you would reach full mana at the exact same time regardless of how much mana you started with (as long as you didn't actually oom midway and actually didn't lose regen for that extra max mana).

Overall this is talking about stuff that generally don't happen in almost any fights where your gear choices could actually matter.

The recent path of discussion reinforces the idea that intellect isn't a good stat until you're running out of mana. Theoretically, it does give you mana, but in practice, if you don't end up using your entire mana bar (which happens very infrequently) it's only use is via Holy Guidance and the Crit, which still makes it worse than just getting Healing or Crit.

Unless you actually make it down to (or very near) zero mana, you haven't used that buffer of 300 mana. Say you have 10,000 mana, spend 2000 then regen back to full. You've spent 2,000, and have 10,000 left (12,000 total). Now add another 1000 max mana and do the same thing. You'll have spent 2000 mana and have 11,000 left (13,000), for a gain of exactly 1000. However, in the scenario of going OOM then regenning to full, the total mana would go from 20,000 to 22,000, doubling the benefit of int. It might sound like a stupid semantical argument to say that any mana you add is effectively the last mana you use, but that is how it works out in practice.

With the 2.4 changes, getting a double innervate to go from empty to full midfight might be plausible simply because priests and druids won't need them. Other than that, there aren't really very many situations where that'll actually happen. As galzohar noted, it's probably more useful to look at multiphase fights as multiple fights, and other ways to fill to full are rare enough that they aren't safe to gear around.

You're going to see innervates going to Mages/Spriests instead of Paladins. Hell, you'd see it go to a Shaman before it goes to a Paladin, since we are still the most mana efficient class.

When you're stacking your raid DPS will generally come first when it comes to buffs. It just increases your ability to kill a boss more than stacking your healers in most cases, since it not only increases dps but also reduces healing needed in one way or another. Not to mention if mages end up actually going arcane, giving them more mana will be a direct, noticeable increase to their DPS - unlike with fire where extra mana simply means you can swap some mana pots/gems to destro pots/flame caps.

On bloodboil you get so much mana back from spiritual attunement mana is really not a problem even with spamming HL on the fel rage target assuming you use your cooldowns properly and pot. If I had a shadow priest there it would've been a joke when it comes to mana. Fel rage topping your mana up is just a "nice extra that happens sometimes". But seriously I'd gear for max HPS for that fight, probably even including haste (at least if I'd have a shadow priest), since HPS is just that important, while going oom and therefore mana efficiency is not so much.

I tryed playing with this chart allitle, but still im not sure what is the best way to go right now.


Sometimes my mana seems to "leak" so fast, essencialy on fights i keep the tank, and do raid healing at same time.

I think i still need alot more mana efficiency, over HP/S right now.
Should i invest in Mp5? or stick with Crit?
My Int seems alittle low too, maybe i should try to stack some more?

I want to ask again for experienced Bt/Hyjal healing pallys: What is the best way to healing? FL spam? Downrank HL spam plus lv11 when the tank receive big hits?
Any good advices of how to improve the healing? In general, i mean, good tatics, and "tricks"...

Should i stick to MT healing more and let the raid healing to the other healers? Cause many times people would have died if i didnt healed them. But sometimes this consumes alot of mana, since the raid healers of my guild sometimes slack alot...

Sticking to MT will give me a low Healing and a high Overhealing on Recount-like addons. This may be a problem since my guild takes this kind of addons as a base to who they think are good or bad at DPS and to some extend to healing too.

Well, sorry again by my bad english.
This are all random throughs, and questions i have right now, thx for the help.


Armory Link http://www.wowarmory.com/character-s...ashi&n=Mirkael

Greetings, I just discovered this forum and am a little disappointed that my spreadsheet is being pointed to without attribution - and an old sheet at that. You can always find the latest here:

Paladin Healing Calculator | World of Warcraft Tools | Curse

At the same time I appreciate people using it - and if you find anything wrong please don't hesitate to post. As for changes or additions. If they look reasonable and I have the time (the big problem) I will be happy to update. I do this spreadsheet as a way for me to understand the game mechanics, to help tune my character, and to understand what is and isn't an item upgrade.

The discussion regarding the trade-offs between the [Libram of Absolute Truth] and the [Libram of Mending] is interesting, but I think wholly unnecessary. The optimal approach is, as some have mentioned, to switch between them depending on the status the LoM buff. It's possible to make these switches with a macro, but even with the 30s window coming in 2.4 I think most of us will find it incredibly difficult to keep up.

Luckily, Holy Librams! will do it for us!

I've actually been using this addon for some time now after the disappointing discovery that Librameister doesn't provide an option to switch librams based on either efficiency or throughput. Rather, it makes static decisions using the following info: 1) BoL on/off, and 2) FoL/HL. Holy Librams! does provide this option, and further considers the status of the mending buff when deciding which libram to equip.

When Holy Librams! is in efficiency mode, I can see a boost of ~3 mp5 depending on my HL density. This promises to be even higher in 2.4 - theoretically up to 8 mp5 simply for installing an addon that is 100% transparent. Pretty good deal!

This addon works best if you have the following librams:

• [Libram of Souls Redeemed]
• [Blessed Book of Nagrand] (or [Libram of Light])
• [Libram of Absolute Truth]
• [Libram of Mending]
• [Libram of the Lightbringer]

It has two modes, throughput mode and efficiency mode:

• Efficiency mode - /ls mp5 - Switches librams to minimize your mana usage (default).
• Throughput mode - /ls hps - Switches librams to maximize your healing per second.

Here's the Flash of Light logic:

And here's the Holy Light logic (roughly):

The code is short and simple enough that anyone should be able to understand it in case you want more detail.

Theory

The maximum theoretical savings for the system outlined above is easy to calculate. Because LoAT heals cancel out, it's purely a question of how many times the Mending buff activates. For each activation you save 17 mana. With the 10s buff, the maximum savings is therefore ((44m - 27m) / (10s / 5s)) = 8.5 mp5.

Realistic predictions are a little harder. Most of us probably don't cast much more than one HL per 10s on average even on HL-friendly fights, so with evenly distributed HLs we would actually see better returns by just sticking with the LoM. Of course, we don't generally get evenly distributed HLs either, heh.

The 2.4 changes are nice because as the duration of the Mending buff goes up, the need for an even distribution of HLs goes down and we approach the theoretical maximum. With the 2.4 stats, that number is ((132m - 84m) / (30s / 5s)) = 8 mp5. It's a bit lower, but we should be able to get closer to it with fewer casts.

Empirical Tests

In case anyone is interested in some real-life numbers for mp5 mode, I've compiled results from 12 separate Vashj fights on a live server (that is, the [Libram of Mending] buff is only active for 10s and the [Libram of Absolute Truth] only discounts 27 mana).

About my methodology: I browsed the log for each attempt (not including OOC time) and noted when the Mending buff (which is conveniently named "Light's Grace" in 2.3 ffs) proc'd either alone (indicating Light's Grace was already up) or within 1-2 seconds of Light's Grace (indicating both were down). I tallied these up in a spreadsheet against my HL casts, my HL percentage, and the fight time.

First of all, this addon's effectiveness is (obviously) highly correlated with the number of HLs you cast. In only 12 attempts it's a little bit difficult to make any precise determinations, but (assuming the ideal trendline goes through 0,0) this chart makes the relationship look pretty linear:


Of course it's not just the number of HLs, but also how bursty you are with them. My HL percentage (and therefore my HL density) was very high in these fights (56% on average), so this addon worked out to be pretty effective. Here are my mp5 numbers for each attempt:


That works out to an average of 3.15 mp5 over what I would have had by sticking purely with the [Libram of Absolute Truth]. I didn't calculate these numbers against a constant [Libram of Mending] because I take that to be somewhat less still (and because it would be friggin hard).

Your mileage may vary, of course.

Get It

2.3: A version compatible with 2.3 is available at the major addon sites: Curse, wowui or WoWInterface.

2.4: If you'd like to try it out on the PTR, please see the Sourceforge repository.

Feel free to let me know what you think!

i think these libram addons should consider adding a popup warning to not cancel on libram switches since it invokes the gcd, most of the time its probably better to let the heal land.

Interesting idea, and I should admit that Holy Librams! will be worse on that count than Librameister because it switches librams more often for the same spells.

I guess I'm not sure I see the problem as too significant, though. Here're the scenarios I can think of:

• Precasting/canceling FoL. Does anyone actually do this? I actually find it hard to cancel a FoL before it goes off without completely losing the benefit of the precasting. Besides, FoL is cheap.
• Precasting/canceling HL. This is common, but I expect most paladins are canceling heals after or near the end of the GCD anyway.
• Aborting to move. Doesn't really matter if the GCD is ticking while you run around, of course.
• Aborting for an emergency cast. This is the one case that seems problematic to me. I have a /stopcast on my BoP, but you're right that the cast won't work if I just switched librams. I don't see any way around this.

In any case, I'm not sure a warning will really help very much.

I'm been looking at this and believe your FoL code is correct for both 2.3 and 2.4. Also, I can't argue with your Holy Light code for 2.3. But for 2.4 I think there is no reason to use Libram of Souls Redeemed for any Holy Light casting. Getting 84 mana back on every Holy Light cast is >> then 120 additional healing from a Holy Light Cast. I understand you may at times want to maximize your HPS but the trade-off of having your heal spell be 10% to 20% less mana cost vs adding 3-5% healing per spell (oh 1-3% HPS) seems so heavily weighted toward mana that there would be very few situations where you'd want LoSR.

Darion

I pretty much agree, but one of the things I didn't like about Librameister was the inability to easily make that choice for myself. On trash, for instance, there's typically no reason not to maximize your throughput. For most boss fights, on the other hand, HL9-11 are almost always plenty of throughput when you compare the potential mana savings from mp5 mode.

It defaults to mp5 mode, though, so your HLs will never equip the LoSR unless you explicitly enable hps mode.

I was looking through the code, one problem I found was that it switches items on SPELLCAST_SENT. The problem with that is with the changes in 2.3 that can be sent even though you are still casting. So, it is typical now for people to spam the spell as the cast is almost done, but if you send the event before the spell actually finishes you can switch to the libram for the next spell before it starts casting. Causing GCD issues, and not receiving the buff from LoM.

You can use SPELLCAST_START, but that triggers a little bit later then SENT due to lag. I made an addon similar to this but I restricted it just to Libram of Mending/Spellsurge so you only get slight delayed (.1-.2 sec) casts every 30-40seconds. I have the option to either put on your FoL item (BBoN) or HL item (LoAT) and you can just choose which is most appropriate before each fight. I also don't like running into possible GCD issues (for stop casted spells) everytime I swap to a different spell. Having it only happen twice every 40 seconds is much more acceptable for me.

Granted this isn't an issue if you don't spam your button near end of casts. But, I find I tend to do that sometimes even when I try not to. Mostly when I NEED to get a heal off, where it will be most detrimental. Also, it is the most effective way that I have found to ensure the shortest cast time (with lag) of all of your spells.

I think this is partly right.

In 2.3, there are three things that can happen when you cast a new spell (Spell2) during a previous spell cast (Spell1):

• During Spell1's GCD, I believe the client aborts the cast. No extra GCD, no SPELLCAST_SENT received.
• After Spell1's GCD, the client starts a new GCD and requests a spellcast from the server. Two possible outcomes:
  • It's too early to queue a new cast - server denies the request, client aborts the GCD, no SPELLCAST_SENT received.
  • Spell2's cast is queued - Spell2's GCD continues, SPELLCAST_SENT is received.

Without a SPELLCAST_SENT this addon doesn't do anything, so I don't take the top two outcomes to be relevant. As for the third, you're right that it's a problem, but the extent of the problem is that you've equipped the wrong libram for one cast. There do not seem to be any GCD issues. It lessens the effectiveness of the addon, but it does not put anyone at risk.

The only GCD issues that I've found were identified by the above poster - I cannot /stopcast heals during a GCD (for BoP, say).

I actually tried this out, but it has different limitations. For one, you still can't /stopcast heals. Second, SPELLCAST_START doesn't provide the target of a spell, so checking for BoL is out. And third, while a FoL doesn't interrupt the mending buff, the event still isn't late enough for a second HL to sense a new mending buff (thereby causing a second, mostly useless mending buff). Overall it seemed that it wasn't worth it.

Honestly, I'm unable to cause any sort of GCD problems, even spamming the button - the timing of my heals does not seem to be affected by this addon.

Were you in combat while testing it? Swapping items don't cause a GCD out of combat. On the ptr I have tested using SENT and spamming and definetly gave me an extra GCD when swapping a Libram early. If you are not on GCD (even if casting a spell) and press a spell it will send SENT.

Yea, it causes LoM Libram to be on for two spells, but I think that is relatively minor compare to excess GCDs. I don't check for BoL either because I don't like be on GCD for almost any spell. In case if I want to switch a quick spell to someone else, like a tank. Also, if you wanted to support BoL, just have a SPELLCAST_SENT method that remembers what was the target of the spell and just use it in the SPELLCAST_STAT.

With regards to your addon Ladenedge, I think it's awesome. I found three things which I'm not sure are intended, or you just didn't think it was necessary to code. I don't have the Lurker libram by the way, and it bugs me no end for passing it all those months ago.

Anyway:

1) In hps mode, without Blessing of Light, when I spam HL it constantly tries to equip [Libram of the Lightbringer] every time i press the button, and gives me the chat message too. This doesn't happen when I spam FoL with or without Blessing of Light, or HL in mp5 mode.

2) In hps mode, I would still like to keep up the mp5 buff from [Libram of Mending] up at all times, especially now that it lasts 30seconds.

3) Finally, in mp5 mode because I don't have the lurker libram, it would be awesome if the addon equipped a more suitable trinket for the extra HL casts in between keeping the mp5 buff up.

I realise adding in options like these would a be a bit of a pain to such a simple addon. Points 2 and 3 are pretty much the same thing, they are both the same problem for someone without the lurker libram. Correcting either one would solve the problem. But anyway keep up the good work

One thing people may be missing is that the advantage of all this Libram swapping is dependent on the healing style you have. I happen to be a Holy Light user - 2/3'rd of spells cast are HL and 1/3 are FoL. Or another way to look at it is 85% of my healing is HL based and 15% is FoL based. So let's look at the math behind swapping for me. This is all 2.4 based and is based on me casting HL spells on average every 6 sec and healing spells on average about every 4 seconds (based on WWS logs for SSC boss fights)

Benefit of swapping:
- 20 Mp5 buff. Why not 22 Mp5? I assume the mods used keep Libram of Mending 22 Mp5 buff up on average for 30 seconds every 33 seconds. Basically getting 90% of the benefit of this buff. All mods rely on swapping happening after the buff goes down. If I cast a HL spell on average every 6 seconds, then on average the next HL spell would be 3 seconds after the buff goes down.
- 3.6% increase in FoL average healing. Why - FoL gets benefit of Blessed Book of Nagrand (79 +healing) if BoL isn't on the target and Libram of Souls Redeemed (60 extra healing) if BoL is active. The math for my +healing (1743) yields 3.6%.

Negatives of swapping:
- -14 Mp5. Lose benefit of -84 mana on one Holy Light every 30 seconds. This translates to -14 Mp5. This almost negates the 20 Mp5 buff from using the Libram of Mending.

Net Positive: 6 Mp5 buff and 3.6% increase in FoL average healing.
Net Negative: Potential GCD issues with interrupting and recasting

I will use the Libram swapping for a bit and see if I can identify real life GCD issues before making the the final decision.

I haven't gone through the thought process yet on FoL spamming but I suspect (no analysis to back it up) that Libram swapping will prove more valuable to that style of healing.

And lastly, I have updated my spreadsheet to add:

- overhealing modifier
- Libram swapping modeling
- Trinket - Seaspray Albatross
- Healing output section that provides Fight HPS and # healing spells cast
- Section that Where your healing is coming from - HL or FoL

Find new spreadsheet here sometime this morning:

Paladin Healing Calculator | World of Warcraft Tools | Curse

Oops. Yeah, you're definitely right that spamming the button provoked a number of spurious SENT events and can extend a GCD past the end of the spell (without starting a new spell).

Lemme see if I can sum up the problem for my own benefit. If one starts the spell too early for it to be queued, the server aborts the spell, but not the libram equip. In that case, the libram's GCD continues where the spell's GCD would have aborted. I guess I haven't seen this problem very much because I tend not to send spellcasts before they can be queued. If you do routinely get your spellcasts aborted (including spammers), or your revert to that behavior in emergencies, working with SENT is probably not a good way to go.

Gah, what a ridiculous bug. I released a new version this morning that fixes this. As for the configuration suggestions, I think I would like to avoid adding much in that way. The source is simple enough that I bet almost anyone could adjust the algorithm for their custom needs (like missing librams, etc.).

Actually, Holy Librams! has an option to set the time, in seconds, less than which the Mending buff will be reapplied. It defaults to 2s (the length of a HL) as a result of 2.3's timings, but under 2.4 it might be reasonable for someone with a sparse HL distribution to set it to 15 or 20s. In your case, it sounds like you would want to bump it up to 3s.

/hl refresh <n> : Refresh Grace of the Naaru when it is under n seconds.