Yes, I ignored IoL procs to keep things simple and because at a first look it doesn't seem like it's going to increase the value of crit by much, but I'll try to think about it.
On a side note, I forgot to account for the +10% crit to HS from IoL, I'll fix that soon.

The answer to this question is simply that based on those assumptions there would not be any benefit to using DL over FoL.

However, I think the assumption that you can "keep from running out of mana" while spamming FoL is quite a wild assumption. That may be doable in a very narrow set of circumstances, but those circumstances would be very rare. Its similar to saying "assume I could turn gravity off, I could then run up hill as fast as running down hill". Probably true but equally useless given that mana is now our constrained resource rather than time.

Elvez, I'm pleased we are moving toward more concrete values for our stats. This has been stated before, but a large part of what is missing in your evaluation of Crit rating is the "mana-free" aspect. As crits aren't nearly as susceptible to overhealing as they were in WotLK, we should look at crit not as a higher HPS stat, but as a higher HPS + HPM stat. I have not done heroic modes, but my experience with the first few raid bosses has left me mana starved more than HPS limited (anecdotal). Is there an easy way to add this value into your weighting?

This has also been stated, but the synergies between Crit and Mastery is also important, although much more difficult to quantify. You have already setup 2 situations to evaluate mastery, one at 50% effective and one at 100% effective. Would modifying the Crit weight to include synergy from Mastery at these two effective levels be possible?

Apparently I edited my post right when you were answering...
Of course what makes PotI really valuable is the healing through the beacon, and since it can crit, you're right in saying it will increase the value of Crit, as PotI is unaffected by mastery.
EDIT: PotI is too important to leave it out of the evaluation, only now I realize that. Fortunately it shouldn't be hard to add it.

That's really what I'm trying to figure out. From all your feedback I'm starting to see more the value of crit from a mana saving standpoint, but still haste is not bad from that perspective too as it allows for more spam of cheaper heals.



For the mana-saving component of crit, it's not hard to estimate its value in term of mana per spell saved. What is hard is comparing output increases with mana savings, as they are two different quantities. I may be wrong on this, but at the moment the only way I can see to convert mana saved to output values is by considering a given fight length (such that X mana saved will allow you to cast Y more spells), but that would probably require a more accurate modelling.
EDIT: I've been thinking about seeing crit as a decrease (on average) on the cost of spells more than an absolute mana saving. This should allow a comparison between output gains provided from other stats and crit, by assuming each spell must cost the same amount of mana, and so accounting for the higher HPM given by crit. I'll try to see if I can sort out the math on that...

That shouldn't be hard, but we'll need to set a fixed value for the mastery value too.
And even though it doesn't look like it's going to bump the value of crit by much, I'll try anyway to add that into the equation.

Sorry, Raphirau, but your formula doesn't work.

At 2 second damage intervals you will get a proc from 2/2 BL at T+0, be unable to proc at T+2, T+4 and T+6, then proc again at T+8 and so on. 1/2 BL has a 50% chance to proc at T+0; if it doesn't it will have a 50% chance to proc at T+2, T+4, T+6 and so on until it does proc. However, if it doesn't proc before T+8, the maximum number of procs you can have in the fight is reduced by 1. This should typically happen once in 16 cycles, so the average gain for having 2/2 would be 1 HoPo every 144 seconds.

For five second intervals, you will always miss your next chance to proc when you do proc. So 2/2 BL will proc at T+0, be unable to proc at T+5, proc at T+10 etc. This time 1/2 BL must proc at either T+0 or T+5 to avoid losing a proc, which is only 75% likely - you will lose 1 proc in 4. Effective CD of BL has raised to 10 seconds, so here, the average gain is 1 HoPo every 40 seconds.

Finally, if damage is being inflicted less frequently than Blessed Life can proc, then it will have a chance to proc every time damage is inflicted. In such a scenario, across the infinite series 2/2 BL will provide twice as many procs as 1/2.

Malleus is right that those numbers are off. As a sanity check, consider the damage every 9 seconds scenario. with 2/2 BL it will proc every time, and in a 500 sec fight, that is 500/9 = 55.6 procs. With 1/2 BL you will average half that (27.8 procs). That is of course the worst case scenario for 1/2 BL (being hit slightly less often than the proc cooldown). Shorter or longer average times between damage mean less difference.

I'd suggest that in most cases this worst case scenario is unlikely: if you are being meleed by a mob you are getting hit more often (and how many HP you generate is the least of your worries), and if you aren't then you are likely looking at intermittent abilities that target random raid members. Most of those don't strike you once every 9 seconds on average.

Meanwhile that point could be put in enlightened judgements. If you judge every cooldown - not always possible obviously - you generate (500/8) * 1334 extra healing on yourself- an extra 83k self healing, and more importantly, 41.7k healing on your beacon target. That's worth a few WoG casts (typically of the order of 15k per WoG, depending on gear obviously). That's around 8 HP's worth of WoG casts. (EDIT: that number is based on the raw healing listed in the talent calculator; with various bonuses factored I get about 5000 per proc with 2 points in enlightened judgements, so about twice the base listed value, meaning more like 83k healing on the tank per point in EJ in this scenario, which is about 16 holy power equivalent of WoG casts).

If you are using the extra holy power for LoD casts instead then obviously it depends on how many people you hit, how much overhealing there was, and all the othe usual confouding factors.

Adding to this...

HPS is not always the best judge of healing in Cata. If you had the chance to heal the entire raid to full at the expense of all of your mana that would be an insane HPS and possibly even a decent HPM. The problem is, you're now useless until you can regen your mana back. HPS, now, seems to be more of a judge of emergency healing and immediate triage than it is a good judge of overall healing.

I've found that at acceptable levels of spirit and haste; I go from 100% -> 50% during burst healing... be it on the raid or on the tank and then healing (triage) with HL / HS and judging during off damage times. My mana ends up back close to 90% before I need to burst again. Toss in DP / AT and I am very mana effiecient and generally am able to push out an insane amount of heals.

I generally find that DL is much much more useful than FoL in most instances because incoming damage is not often going to kill someone if they don't get a heal NOW. (I have not experienced heroic raids yet, so this is ignoring that aspect for now). The only exemption I have found is Chimaeron, which requires procs from EG or HS crits to keep you from needing FoL. FoL, as I see it currently, is the best heal to keep someone alive when you know that they're about to get hit or could get hit and their health is low. One or two FoL and then a DL is probably the best way to top someone off in those situations... but it burns mana. DL will heal for more than FoL, it just takes longer to get it to the target.

This leads to a balance between HPM and HPS; which in the vast majority of cases you'll want to lean towards HPM to provide longevity. At higher gear levels, FoL and DL might be used more frequently due to higher regen... but the same claim could be said about HL if you forgo the spirit and get instead haste and additional SP. HL, for me, heals for a very good amount and I am essentially always casting it even if damage is minimal. More spell casts leads to more procs from Daybreak which leads to more HP. This is in sharp contradiction to another holy paladin I know whom constantly uses FoL. He might be saving the tanks more when there is spike damage... but he's also pushing out FoL when there is no imminent need of healing. This wastes his mana. In the end, our healing is drastically different as I am able to constantly throw out heals to the raid for a much longer period of time. (Which also means more uses of HR since mana is not as much of an issues).

It comes down to play-style and raid role. If you're healing the tank you want to keep him topped off but not if it means you can't heal later on when your mana is gone and damage is still incoming. It's essentially come down to what blizzard wanted, you have to make a choice, and the difference between a good and a bad healer has never been more apparent.

It does proc from Light of Dawn.

Not to be too pedantic, but this calculation is also inaccurate (however, it does not qualitatively change the analysis, which is that 1/2 Blessed Life is often more valuable than half of 2/2 Blessed Life). It overestimates the value of 1/2 Blessed Life somewhat because you do in fact lose something when you miss even one proc (in particular, a fraction of holy power). In the 2s interval example, every missed roll is the loss of 25% of a Holy Power (rather than just losing 1 if you happen to fail 4 in a row).

The correct formula for the expected holy power fraction in the infinite case should be N/(1+N) where N is the number of ticks available per 8s interval. Intuitively, this is because the number of expected missed ticks before a successful proc is 1 (given 50% miss rate; it's possible to go through the actual probability expansion if so inclined). In practice though this figure is highly dependent on the actual distribution , usually in favor of 1/2; if for example you get 3 rapid bursts in 8s, in practice you only lose out if all three fail (so 87.5%) rather than the 75% predicted by the infinite case. This is because you can't actually lose fractional holy power.

Has anyone given any thought to using the Ember Skyflare Diamond rather than the Ember Shadowspirit Diamond?
The difference, of course is 33 Int (54 - 21), but since it requires 2 yellow gems, we have to sacrifice 40 int to activate it.

Rather than using 2 orange gems to activate the bonus, we could use two 40 int gems in their place, which would give us an extra 20 intellect per gem. This means after taking the lower int of the skyflare into account, we'd be netting a 7 intellect gain due to an easier requirement.

Am I missing a hotfix in regards to requirements that isn't displayed in game yet?

EDIT: I was looking at it the wrong way. It's basically choosing 7 int vs. 40 haste. Int is good, but not 6x better than haste (before socket bonuses are considered)

I've worked out the blessed life thing to my own satisfaction, both analytically and via numerical simulation (getting consistent results in both cases). I've expressed the results in terms of Holy Power gained per minute for a constant attack speed:



x axis is attack speed (seconds between attacks). y axis is holy power generated per minute. The blue line is for 1/2 Blessed Life, the red line for 2/2 Blessed Life.

Perhaps a better graph is for 1/2 Blessed Life and the difference between 1/2 and 2/2 BL. This shows the value of the 1st point, and the value of the 2nd point.



The blue and red lines overlap for attack speeds slower than 8s. The blue line is the value of the first point you put in to BL. The red line is the value of the 2nd point. There are no great surprises here: When you get hit rarely, both points are equally valuable (i.e. you get twice as meny procs with 2/2 as you do with 1/2). For very fast attack speeds hitting you, the 1st point is valuable, and the 2nd point is largely useless.

Of course, you are almost never being hit at a constant attack speed throughout a fight. You might be randomly targetted with abilities once in a while (rarely more often than once every 8 seconds though) and have the occasional moment of having an add as your personal friend until it is tanked / CCed / killed / kited. It's up to individuals to figure out how often they are getting hits that might proc BL in practise, and judge how many holy power per minute they need to make it worth putting the points in blessed life rather than in some other talent.