Int might not be as good for Alliance paladins, but I am not going to ignore Arcane Torrent for myself since I am a Blood Elf. I broke it up in the calculations so if you didn't have it you'll know Int is 4.7mp5 worse for you in this example.
No, it's 4.7 MP5 better for you. That may seem like a tautology, but it isn't. This is the Paladin thread, not the Belfadin thread. Just as if you were giving advice on reaching the hit cap you would not base your numbers on a Draenei, when you're advising all Paladins on how they should gem you base it on the abilities possessed by all Paladins. Racials is extra.
I got my stats for Replenishment uptime by examining individual boss fights on a number of WWS reports. I selected for 25-man speed runs as those groups are more likely to have multiple Replenishers. After that, it was simply a matter of dividing the number of ticks by the duration of the fight. That may not be a strictly accurate measure of uptime for whatever reason - lost ticks to refreshing, for example - but it is a perfect measure of effective uptime. If I fight for 300 seconds and get 240 ticks, then I effectively only had 80% uptime even if I had the buff for the full five minutes.
As for ignoring SP and crit from Int: it is inevitable that at some point returns from MP5 will overtake increased base mana. Replenishment has moved that point further away from zero, but it's going to happen. Over the timescale we are discussing, ignoring a cumulative gain below 1% is moving that goalpost by at most 3-5 seconds. It's not significant enough to make a difference.
No, it's 4.7 MP5 better for you. That may seem like a tautology, but it isn't. This is the Paladin thread, not the Belfadin thread. Just as if you were giving advice on reaching the hit cap you would not base your numbers on a Draenei, when you're advising all Paladins on how they should gem you base it on the abilities possessed by all Paladins. Racials is extra.
That is why it is broken up, so you can see how much mp5 you get if you are a Blood Elf or if you are not. Its not called the Allianceadin thread either, you can't just ignore something in the general case that 50% of the players have.
I got my stats for Replenishment uptime by examining individual boss fights on a number of WWS reports. I selected for 25-man speed runs as those groups are more likely to have multiple Replenishers. After that, it was simply a matter of dividing the number of ticks by the duration of the fight. That may not be a strictly accurate measure of uptime for whatever reason - lost ticks to refreshing, for example - but it is a perfect measure of effective uptime. If I fight for 300 seconds and get 240 ticks, then I effectively only had 80% uptime even if I had the buff for the full five minutes.
Looking through WWS reports, replenishment uptime seemed to vary a lot. Both from fight to fight and guild to guild. Some have near 100% uptimes and some below 80%. But 90% by far seemed like the average uptime, so that is the number I am going to use
As for ignoring SP and crit from Int: it is inevitable that at some point returns from MP5 will overtake increased base mana. Replenishment has moved that point further away from zero, but it's going to happen. Over the timescale we are discussing, ignoring a cumulative gain below 1% is moving that goalpost by at most 3-5 seconds. It's not significant enough to make a difference.
If the crit from Intellect is that negligible that it isn't even worth calculating when talking about Int, then why would you ever get any crit rating on your gear? Since Crit Rating only provides twice as much Crit as Intellect does per item point.
Also, one major thing I forgot to calculate was Divine Plea. Which with 100int gives you 474 extra mana per use, or 39.5mp5 if used on CD. Or even if you use it only every 2 mins, which makes it 19.8mp5, it is just as good replenishment.
From 100 Intellect you get:
1897 mana at the start of the fight.
19.8 to 39.5 mp5 from Divine Plea
21.3mp5 from Replenishment on average
4.7mp5 from Arcane Torrent
25.3 Spell Power
0.825% Crit Rating
Which ends up being a lot of mana, and extra healing done.
Also, one major thing I forgot to calculate was Divine Plea. Which with 100int gives you 474 extra mana per use, or 39.5mp5 if used on CD. Or even if you use it only every 2 mins, which makes it 19.8mp5, it is just as good replenishment.
I'd assume that DP would be used on CD or as near to CD as possible every time. That would mean for Paladins at lvl71+, the mana boost from DP alone on 100 Int would match the 40 MP5 from five equivalent gems. Combined with Replenishment even at 80% uptime, it's strictly better. I yield the point.
I was wondering lately about weapon enchants. Healing Sunwell has become rather easy, to put it mildly, and the overhealing caused by holy shock/holy light combination seems currently monstruous. Since INT affects all those stats named in previous posts (crit, mp5 from replenishment, more base healing and from the holy talent, more divine plea mana in the future, etc.), would not the +30 INT to weapon become viable for now, at least until WOTLK?
I have been doing some reading up and research as to the discussion on the last 2 pages of INT vs Spell damage vs Crit. Seems as if the new way to go is towards Int, and better yet towards Int-Mp5.
Sources such as Maxdps.com have now put the INT-Mp5 gem at the top of the HPS cycle and have also stated that any enchant w/+INT is more benificial to the Holy pally.
The tough decision is, currentlly are the enchants that are out there powerful enough to override the +sell damage (old +healing) that we get. The post right before me touched on it by asking the same, I don't know if I want to get ride of my expnsive +healing enchant to try it out and find that it did little to help or actually stayed the same. Is major int on a weapon vs major healing now better for the Holy Pally?
I spent close to 300g in gems last nite playing with the +Int/Mp5 vs the +spell damage and seen hardly no difference at all, the Int/mp5 seemed to crit more but that may just have been me being tired too.
I would like to know what others have done to test this theroy out and...if Maxdps isn't a reliable source please don't just dicredit them, explain to me why you have noticed the difference.
Thanks.
Dryad, well if healing Sunwell is easy, then it doesn't really matter how you gem really. Also, if overhealing by Holy Shock/Holy Light by a lot, then Int probably won't help you. Since, the main benefit of Int is that it lets you cast more Holy Lights before going OOM.
zumhug, I don't trust anything from maxdps.com, it might on occasion be correct but there have been far too many times when they aren't. It probably is not worth blowing your money to change all your gems around for 2 weeks of heavily nerfed content. I am not sure yet if I am going to gem more for Spell Power or Int in WotLK yet, I will have to do the raid encounters to see if I need more burst healing/stream hps (Spell Power) or if I want to be able to Holy Light more freely and often, but for less (Intellect).
Has anyone done any testing to verify the new Spell Coefficients? I have seen serveral places mention ~1.8 or 1.9 scaling on the coeffs but nothing conclusive with proof. Using the Wowwiki Item costs and doing a ratio between SP and HP gives 66/35, or ~1.89.
Some initial testing is showing me at least a factor of 2 (for paladins). The Blessing of Light reworking is likey the cause, but I was under the assumption they would just add that to base spell numbers, as that is essentially what the buff did. Testing shows 1:1 with tooltip numbers at 0 spell power. Also, FoL and 2.5s HL still appear to have the same coeff scaling per second, meaning that whatever changed has the same relative effect on both spells. (which seems to rule out a hidden flat bonus)
Some druids doing testing on the Beta forums appears to confirm a ~1.8-1.9 scaling on the base Coefs, but from what I can see, paladin heals (FoL and HL so far) are showing around 1.25x more then this, for aprox. 2.35x the old coeffs.
Initial results are showing ~1.01x for FoL and ~1.69x for HL coefficients, if you guys have narrowed it down it would be great, but I will keep working in the meantime.
It is 100% for FoL, and 166% from HL, but remember you need to make sure to account for Healing light. Healing got converted into Spell Power by dividing it by 1.8. But the spell coefficients seemed to got changed at a different coefficient, multiplied by 7/3.
I am 100% certain it is not exactly 1.0 for the FoL coefficient. This is easily verifiable. At 602 spell power, 0/0/0 spec, FoL had a healing range from 1206-1277, subtracting the 600-670 from FoL7 @ 70 gives 606-607 bonus from 602 +spellpower.
At 1075 sp, same conditions, FoL7 range of heal was 1684-1755. Bonus of 1084-1085.
I will test some more to make sure there is nothing out of whack with enchants or gems scaling oddly.
Also I did a series at 924sp, giving 932-933 bonus healing.
Edit2: Another test with 258sp from items that had no enchants and no gems. Bonus healing was 261-262.
Edit3: It seems that the elegant form of the FoL coefficient should be something like this: 1.5/3.5 * 66/35 * 1.25 = ~1.01. This would imply blizzard gave us a flat 25% bonus scaling on + spellpower. The same rule applied to HL11 gives similar results, 2.5/3.5 * 66/35 * 1.25 = ~1.68, which coincides with my HL11 results, but those are much slower because of mana cost and a greater range of healing.
Most tests were with a fairly small sample, less then 50 casts for most of them. The key is using the range of healing to eliminate the randomness of the middle ground. With low sp gear on, my mana pool was low so that limited the casts, however the ranges do seem to line up fairly nicely, baring being out maybe 1-2 healing for those outer values.
I did some reverse testing, and the +25% seems to line up decently, but there are some minimun heals that are out of the expected range, some of them can be attributed to rounding, but there were 2 that were a few points out of the range. It could also be a simplification on blizzard's part in the 66/35 ratio, a flat 1.88 might be a closer fit.
Edit: This is the thread with the druids doing testing. I can't seem to find any other discussions on the new coefficients for any classes.
Edit2: It just occured to me that the old +healing gems and enchants (mainly enchants) might not be a whole number of spell power. My test with non enchanted gear was a perfect fit for the expected range of healing.
So as a new Paladin. I just got the T6 Belt, T6 Legs and the Epic BP from Heroic MGT.
In terms of gemming, after reading this thread, I'm leaning towards 10 int in the heroic MGT for all 3 slots, the 6sp/5 int for the belt and legs (with a third 6sp/5 int gem in my helm). Which will activate the Ember Skyfire Diamond for 14 ap and 3% int.
Does that make sense? That gives me heavily scaling int between Kings, the Meta and Divine Intellect. Which adds SP/Crit/Manapool and Mana Regen based on my reading here.
Quick answer? It likey won't matter too much at this point how you gem it. Int for more staying power (or mp5 if you lack replensishment and other mana sources), Spell power for steady throughput and efficiency scaling, crit for a bit of both worlds, or Haste for maximal throughput at the cost of scaling power.
Most tests were with a fairly small sample, less then 50 casts for most of them. The key is using the range of healing to eliminate the randomness of the middle ground. With low sp gear on, my mana pool was low so that limited the casts, however the ranges do seem to line up fairly nicely, baring being out maybe 1-2 healing for those outer values.
I did some reverse testing, and the +25% seems to line up decently, but there are some minimun heals that are out of the expected range, some of them can be attributed to rounding, but there were 2 that were a few points out of the range. It could also be a simplification on blizzard's part in the 66/35 ratio, a flat 1.88 might be a closer fit.
Edit: This is the threat with the druids doing testing. I can't seem to find any other discussions on the new coefficients for any classes.
50 casts is wayyy too little for something with a range of 70. Though I agree it might not be exactly 1, and 1.66. I am going to run some more tests.
I think it is safe to assume though that whatever they end up being is that HL gets 2/3 more then FoL.
50 casts is wayyy too little for something with a range of 70. Though I agree it might not be exactly 1, and 1.66. I am going to run some more tests.
I think it is safe to assume though that whatever they end up being is that HL gets 2/3 more then FoL.
Every one of my data sets had a final healing range of 71, which should constitute the entire range of healing possible. I did find it odd that it was 71 instead of 70...
Ok, testing 353 sp from gear that did NOT have the old +healing enchants. My predicted range of healing is 696.6-1026.6. Over 200 casts did not result in a 696 or a 1027 heal. I am thus led to believe, A) Paladins have 25% more scaling on +heal then other classes, and B) The old +healing enchants (and gems) have been converted to a decimal number that is rounded on the ui display. My gear had 1 +22 gem, and 1 +11/2 gem. This would result in a 0.5 sp difference, which would result in ~0.505 difference in the end heal. This still should suggest that a 656 heal should have been present.
Okay, so I spammed 191 FoLs on myself and found that the coefficient for FoL is exactly one. But one interesting thing I found out is that the FoL Spell Power libram doesn't gain the benefit of Healing Light.
I had 1045 Spell Power, +79 FoL spell power libram, and +12% FoL from Healing Light. From all my results I got a minimum of 1921, a maximum of 1999, and an average of 1964.
Which is exactly the values that I got. I will do some more testing with various levels of spell power, w/o libram, and next time I go ret w/o Healing Light.
Okay, so I spammed 191 FoLs on myself and found that the coefficient for FoL is exactly one. But one interesting thing I found out is that the FoL Spell Power libram doesn't gain the benefit of Healing Light.
I had 1045 Spell Power, +79 FoL spell power libram, and +12% FoL from Healing Light. From all my results I got a minimum of 1921, a maximum of 1999, and an average of 1964.
Which is exactly the values that I got. I will do some more testing with various levels of spell power, w/o libram, and next time I go ret w/o Healing Light.
Yea, I tested w/o the libram of SR as its mechanics are unknown atm and just convolutes the testing.
600 + 353 = 953
670 + 353 = 1023
My range was 957 - 1026 w/o libram 0/0/0 spec.
Gotta heal a ZA quick, more testing after.
Edit2: My one test might have had a libram on so I removed it.
Hmm, the plot thickens. So I did 243 test with 743 Spell Power, 12% Healing Light, and no Libram. I got a coefficient of 1.5/3.5*66/35*1.25 as well.
My tested minimum was 1523 and my maximum was 1602. My expected 1,523.97 minimum, and 1,602.37 maximum. It also seems that no matter how close it is the the next number, they always get rounded down.
Now to figure out how the libram gets applied. The closest I could find with my first results was that the Libram doesn't benefit from the 1.25 bonus, but that was off by 4 healing.
My tested minimum was 1523 and my maximum was 1602. My expected 1,523.97 minimum, and 1,602.37 maximum. It also seems that no matter how close it is the the next number, they always get rounded down.
This makes sense. If Flash has a range of 600-670, then the base amount healed is probably coded as INT(600+(RND(1)*71)).
This makes sense. If Flash has a range of 600-670, then the base amount healed is probably coded as INT(600+(RND(1)*71)).
Wouldn't that give 600-671 as the range? Also Since FoL7 is learned at 66, the healing ammount scales upwards. WoWhead lists it has 588 to 658, but ingame it shows 600-670. Now where did I see that discusison on spell level scaling again...
In a lot of languages Rand(1) will produce a random number that falls between 0<= x < 1, which means that it can never be 1. Also casting a float to an int normally just truncates the decimal (rounds down). So say if Rand(1) returns .99999999999, multiplied by 71 is 70.9999999999, add 600 to become 670.999999999999, and when cast to an int it becomes 670.
Ya, I wish I knew of some resource on spell scaling, ideally a formula, or at worst a table.
I am in the process of making a new first page post for this thread. That I'll send Zurm, thread starter and former guildmate, to update the first post with. It should help with a lot with useless questions, since the first post is incredibly out of date because Zurm went Ret. So if there if you have any suggestions, or a particular topic you want to see covered, please let me know.
Also, do people think it might be better to make a new thread entirely to start fresh for WotLK? I think it will be better, since it will make it easier for newer people looking at older posts. So they don't get confused with all the many different iterations of our abilities/talents during beta, and so they can easily see where new relevant WotLK discussion starts.
Also, do people think it might be better to make a new thread entirely to start fresh for WotLK?
This.
You don't want to confuse a new Healadin with a thread full of outdated information. Since you are writing it, you might as well start the post in a few days.
As a relatively new Healadin - new to the 70 level content anyhow - I agree a new thread would be nice. By all means link to this one, but I've spent a good bit of time meandering through this thread only to realize much of it doesn't apply anymore. WOTLK changes so much, it's a very different experience now.
10/26/08, 9:22 PM
Similar Threads
