You will get random extra or less ticks at random points in the haste range its true. At the moment HST and HTT are bugged and we have no information about any hotfixes I'm afraid.

Right now, I'd say it's more a question of how much extra mana you can afford to spend and if you feel you need more reliable throughput during bursts. It's definitely not something that's gonna make or break you and I'm seeing people on the top of WoL reforging either way.

As long as HTT and HST are displaying abnormal behavior there is no reason to reforge to 3764 (3306 for goblins). While any haste rating will not actively harm your character, at this point you are better off allocating these points to either mastery or crit. One of the problem that arises is that, the more we progress the better the gear we get thus making it very hard to reforge haste down to 871(441 for goblins). This leaves us with :
1. The choice of reforging away all haste to mastery and crit and work with whatever we have left.
2. If we are left with too much haste after reforging we can opt to forgo Ancestral Swiftness and talent into Elemental Mastery instead. Remember for this option we will still need to keep track of haste cap for Riptide, Healing rain and Earthliving Weapon which is 3039 ( 2588 for goblins).

Note :A lot of people are asking me whether it is wise to experiment with breakpoints that randomly give you that extra tick. An example is when haste is at 20.35% ( 8649 with no raid or personal auras, 6214 with one aura, 3894 with both auras). It is not advisable however to do so mainly because we are not aware of how or when this bug occurs making it both unpredictable and unstable. We can never know for sure that the example of 20.35% haste will always produce 8 ticks of HTT.

As I couldn't find a good way to easily determine the effectiveness of mastery for any given fight, I've written a small addon for myself that will print out the average mastery multiplier after a fight. The way it works is that every time you land a heal it will estimate the target's health percentage before the heal landed and then use the amount healed as weight, so that bigger heals have a bigger effect on the final value. If the addon prints 0.42 at the end of a fight and you have 50% mastery, all your mastery together added about 21% healing on average. The addon should provide a very good estimate to how strong mastery is for a fight, however it is not a perfect measure.

Anyway, as the question regarding mastery came up in the simple questions thread and it didn't seem like there was actually a whole lot of data behind the 50% recommendation, I figured others might be interested in an addon like this as well. You can find it here on curseforge.

You can directly compare the "Mastery Effect" number to the numbers given in the main thread's mastery section. If the effect of mastery is say 0.21, then the average HP of your heals' targets is 1 - 0.21 = 79%. If you are mainly AoE healing, then mastery will provide more throughput than crit over the course of the entire fight if the effect of mastery is above 0.33. As this comparison ignores both how mastery may still be stronger and more reliable during high damage parts of a fight and how crit will likely still win out in an HPM comparison, I still wouldn't recommend making this the only factor by which you decide what to go for.

This is absolutely true. One of the reasons I decided to write the addon was that I felt that even with all the information provided in this thread and around the internet, it was still very difficult to optimize stats for my own raid composition and progression.

Hello,

First of all thank you for this great rsham guide, it's very usefull. I post here because I'm a little bite lost and seeking for advice with the HTT and HST bugs. Til now, I was raiding with the 3k3 hast cap (I'm gob), more or less 50% mastery approx 10k spi and all the rest reforged in crit.

Now that I saw your guide, I'm reconsidering my options. I dropped out AS for EM and set my haste lvl at 2588 as you said it 2 or 3 posts earlier (is this correct or did I misunderstood smthg?). The reason behind that is that I can't reforge down my haste below 2k and it's just a waste to aim for the 441 cap. After that I did reforge all my items into crit, setting my mastery at 50-51%.

Well, you can see all that on my armory profile I guess.

It will still be cool if I could have your advice, considering my gear to obtain the best throughput possible. I'm raiding 25 men and I must say that I'm a little bit lost on how to optimize myself at best for raiding atm.

Thank you

A few days ago I updated the Haste section of the guide to reflect all new changes in red with recommendations for reforging. Should be very clear, if not feel free to comment or pm me for more info. Also look at old comments if you are interested in the timeline of the issue, however the thread will now reflect all new info as accurately as possible with the data we have in our hands. I am posting below a table with my findings which can also be found here. I will update the spreadsheet with more data I have in hand however the most important are the ones I am including below. In the table buffs that are used are also represented, although it shouldn't matter for the result as the actual haste percentage is what is essential.

		Buffs		Ticks
Haste	None	5%	10%	HTT	HST	Latency
0.00%	●			6	8	95
5.00%		●		6	9	89
9.01%		●		6	9	88
10.05%		●		6	9	105
10.20%		●		6	9	117
10.58%		●		6	9	89
11.27%		●		7	9	89
13.32%		●		7	9	95
14.03%		●		7	9	84
14.97%		●		7	9	84
20.02%			●	7	9	95
20.72%			●	7	10	89
21.30%			●	7	10	84
22.19%			●	7	10	51
23.16%			●	7	10	84
25.56%			●	7	10	51
27.28%			●	7	10	84
29.67%			●	7	10	50
30.00%			●	8	10	94
30.00%			●	8	10	50
33.43%	●			8	10	84
35.10%	●			8	11	51
43.20%		●		8	11	96
45.11%		●		8	11	53
45.74%			●	8	11	121
47.58%			●	9	12	117
53.03%		●		9	12	51
57.88%			●	9	12	84

What is important to notice are the breakpoints where ticks are increasing or for that matter breakpoints that they should increase but they don't due to latency. In red we can see the points where we should be getting a new tick but we didn't and in green we can see the points were we should get a tick and we did. These occurrences only take place very close to the actual breakpoints (1%-5% difference) and they have to do with server load and our own latency. They also seem to be working differently for HST and HTT. Whereas these two totems in theory share the same haste breakpoints, our observed values place HST very close to its actual breakpoints and subjective mostly to our own latency whereas HTT seems to suffer from both server and our own load. You should not look at these results as a way to introduce special reforge points for HTT and HST. The percentages are subject to change depending on your own server load and your own latency. These values can only be used as a reference point for an example of reverse calculating of server latency in an attempt to prove the theoretical part of the problem. At this point it is safe to completely disregard any haste breakpoints concerning our totems and reforge as if they didn't exist.

The way we calculate HR and HTT haste breakpoints is different which is why they cannot share the same breakpoints. HR is a spell with adjusted duration while HTT and HST have their durations fixed and won't change according to our haste levels. HST and HTT follow the same method of calculation and due to their fixed values, if the problem we are facing with didn't exist they would follow the same breakpoints for different ticks. Note also that what I have stated in the guide is that there isn't a problem in the formula meaning that the way of calculating them is correct so the problem lies elsewhere.
While HTT shows that it follows HR breakpoints in fact it doesn't, the only reason it appears that it does in certain breakpoints is due to server load adding from 200ms up to 1000ms to its duration. The important thing to note here is that world latency and home latency changes over time and location. One place in the world, or one instance, will have a different load, and a different latency value. It also changes over time, depending on what's going on in the instance mentioned, and the amount of people in the instance mentioned. Your personal world MS and home MS will also spike and drop, one second it can be 40 ms, and the other it can be 120 ms. This is in every respect, impossible to predict. Consider also that when you are conducting experiments you are doing so in a safe "stable" environment. While in a raid instance expect these results to already be different. Chasing the magic value, is ultimately futile, because the moment you drop your HTT or HST, you can have low latency, or high latency, server or playerside. Aiming for a high haste value to offset this, and "bank" on being above the required amount of haste is a gamble that can not be recommended.

In game personal testings are not a reliable way of calculating haste breakpoints when server load and your own latency are involved. In fact this is exactly what I am trying to discourage people to do, go on a ghost hunt for breakpoints that are fluid and will be different for every single player due to the reasons described above. You should not base your reforging on luck and in the hopes of getting another tick when you cannot be absolutely sure of when this will occur. My testings and your testings are unique for your character and your server and the fact they somewhat coincide is irrelevant, someone else's testings with different latency and different server load will show different breakpoints. The only reason we conduct this experiment is to prove that theoretical results differ from practical, not to actually calculate new haste breakpoints based on personal findings and assume they will be the same for the whole playerbase.

Concerning reforging.The reasons above explain why we cannot calculate a safe percentage of haste which another tick is guaranteed which will be consistent for every single player in game. While it would be valuable to be able to do so I cannot recommend reforging based on the gamble that is server load and our own latency. With results interchanging in between these factors it is not possible to find a safe spot for the entirety of the resto shamans out there. In fact I would be entirely unwilling to do so even for myself as well as for my own servers resto shamans unless there was a way to guarantee I will always have the same server and home latency.
The tables below present with the percentages of hypothetical haste breakpoints based on various values of ms the server will add to our totems.

HST	# Ticks	Breakpoints	+200ms	+300ms	+400ms	+500ms	+600ms
	1	6.64%	5.24%	4.58%	3.87%	3.23%	2.54%
	2	20.01%	18.45%	17.61%	16.86%	16.11%	15.37%
	3	33.29%	31.54%	30.68%	29.83%	28.99%	28.16%
	4	46.68%	44.77%	43.83%	48.81%	41.89%	40.99%
	5	59.94%	57.92%	56.80%	55.82%	54.86%	53.79%

HTT	# Ticks	Breakpoints	+200ms	+300ms	+400ms	+500ms	+600ms	700ms	770ms	950ms
	1	20.01%	17.61%	16.52%	15.37%	14.25%	13.22%	12.14%	11.39%	9.56%
	2	40.01%	37.22%	35.92%	34.63%	33.29%	32.06%	30.85%	30.00%	27.84%
	3	59.94%	56.80%	55.34%	53.79%	52.38%	50.89%	49.53%	48.53%	46.15%
	4	79.94%	76.44%	74.75%	73.09%	71.45%	69.85%	68.28%	67.15%	64.41%
	5	99.90%	95.98%	94.08%	92.22%	90.39%	88.59%	86.83%	85.61%	82.57%

The values marked with blue reflect the values of my testings above. We can clearly see that HTT and HR almost follow the same breakpoints but not quite. In fact while HR first breakpoint is at 9.98% HTT first is at around 11.39% with an added of around 770ms. For the second breakpoint HR and HTT coincide and HTT has again 770ms added although they are both at 30% (We cannot be absolutely sure that the number for HTT wasn't below that ). For the third breakpoint HR is at 49.98% whereas HTT is at around 46.15% with an 950ms. We can clearly see that there isn't a consistent number of ms added to HTT at every given time so trying to reforge to a higher percentage in order to get that extra tick is not advisable.

As a personal note, whoever wants to take the gamble they are more than welcome to do so, but I cannot recommend it in the guide. What I can do is present with the data needed in order for players to be informed, the decision will always rest within each player.

Remember that as of 5.1 Ancestral Swiftness now has a 1.5 min cooldown, up from 1 min. This unfortunately devalues this talent even more if we don't need it's passive haste. If you feel you need that extra CD of course you should choose this talent. Raid survivability will always be valued above our own numbers however in this guide HPS plays a huge role. My recommendations always aim for best HPS (with a few choice exceptions) however this shouldn't mean that this is the only route to go. As always take any information you need then make an informed decision for your class, your healing assignment and your raid.

I have some questions to submit:


Haste break tiers
While terribly depressing and disappointing, Therya's testing have been very useful, thanks again for finding out and sharing.
I'm not sure I fully understand though, why does this thing happen only to HST and HTT but not to other spells usually subject to haste/ticks tiers? Because they are not spells being cast but "totems" (i.e. like NPCs?) physically casting the spells?
From Therya's post #71 I didn't completely understand if there is an approximate way to calculate the excess of haste you're supposted to be adding, given a certain amount of average latency. Altough I did get his general suggestion is not to aim for those tiers at all and just stop after 871.
Also, if I can add my opinion on the board, Ancestral Swiftness is still a very useful cooldown to use at specific times (despite the fact its cooldown is now higher).
Also about haste, its relevance beyond the 871 tier may be limited, but it's still there. It gives you a marginal edge on speedcasting, gcd, and also faster manareturns on fights where you can make use of Telluric Currents (faster casting = faster mana return, no?)


Glyph of Water Shield
While of course still subject to personal playstyle, as Therya already said, it would be good if in the future that general guideline could be expanded to include HC fights as well. For some of those there's no difference from the normal, but I figure at least a few HC would benefit from the glyph way more than the normal version.
Another question, why isn't the glyph suggested for Garalon? The constant nature damage from Pheromones would suggest a high activation rate of the water shield orbs, unless for some reason that kind of damage doesn't proc it? In which case I wonder why, many other aoe abilities trigger Water Shield, no?


Ascendance
Anybody knows of a detailed test somewhere explaining exactly how this buff works? How the spells get "duplicated", how much time it takes (I always got the feeling it takes some time), how the target of the duplicated effects are smartly chosen among raid members, the approximate radius of the effect, how does it affect AoE spells. For example we know it works on HR "ticks", but what about CH jumps?
I'm trying to understand if it would be worth to use specific rotation of spells to maximize the throughput of the spell, or if you should just do smart-healing as usual.

Thanks for your patience in answering me.
Anybody tested if 5.1 changed anything about it? Not in patch notes, but you never know of course. Actually it makes me wonder, has any blue ever acknowledged this issue or said anything about it?


I've been using it "efficiently" on a lot of fights (10men). Feng in phase 1-2, Spirit Kings, a bit in Will of the Emperor and a lot in Elegon, helped me up a lot there.
It's ironic since I used to hate it in Cata, where it was undoubtely much better.
Still, your point of view gave me a lot to think about, thanks


Oh me neither, but I meant another part of the debuff, sorry if I wasn't clear enough. Allow me to quote Icyveins

"1.While Pheromones is active (so, always), it will constantly deal a small, yet increasing amount of Nature damage to the entire raid, every 2 seconds."

Unless this kind of damage doesn't proc Water Shield (since I didn't test it myself), this fight should be one of the best for the glyph?
If it doesn't proc it I'd like to know why, since there are other AoE abilities that do proc Water Shield just fine.


Oh but I had checked that already (thanks btw! ).
I just wanted a better insight on the buff to understand if there is any valid reason to change our rotation during Ascendance or if we should just play as normal and enjoy the additional healing we get from it.