I'm having an issue with RAWR. Couldn't find an answer on their forums, but when I go to edit TIAJ in completely deletes it from the list. Anyone have any idea what could be going wrong?

The issue is. It doesn't seem to work out with those values. Wheras the adjusted values do work (better).

Both Rawr.Ret and Rawr.Healadin adjusted health and mana by the same amounts (-180 and -280 respectively). Rawr.ProtPaladin didn't, but from experiments I've been doing that seems to be an error (or protPaladin model simply assumes base stats to be correct). So in fact the base health/mana does indeed need to be lower. Wowwiki seems to be wrong.
For all we know


On my human paladin, unbuffed (but talented), according to the ingame character panel
1811 Sta
129 Intellect
with those values...
6934 + 1811*10 = 25044
4394 + 129*15 = 6329

Actual ingame values:
health 24864 : difference = 180
mana 6049 : difference = 280

Proof given ?

Neraya, there is a simple counterproof. Take the cost of your spells, divide by .29 (In the case of Holy Light, different spells have different percentages) and you find exactly 4934 as your base mana. It has been noted for awhile that 18.7 int is for some reason disregarded for mana pool generation, just as 18 stamina is also disregarded. I have no idea why this is the case, but it is the case. Just mouseover the stamina portion of your character pane and you will see it clearly.

I have had this phenomenon reflected in my spreadsheet for most of wrath at the very least (not sure when I initially entered it, but it has been a long time) so it isn't new. The implementation is certainly bizarre though.

I'm just trying to make sense of why that -180/-280 was there in the calculations.
Changing the base health/mana in the character panel is probably not a good idea then as it would have an impact on any calculations based on base health/mana.

So changing the health/mana amounts is thus a bad idea. I should probably just reduce character sta/intel by those amounts before converting them into health/mana (?) like it was before.

Edit: I actually found the explanation for it, and there's no 18.7 at all involved. In fact, 18.7 is even incorrect as it'll introduce a tiny bug.

The first 20 points of Sta will only yield 1 health each, instead of the usual 10. So for the first 20 Sta you only get 20 HP instead of the expected 20x10 (=200). Hence you get 180 less health than expected.

Similarly, the first 20 points of Intel only give you 1 mana each. So the first 20 Int give you 20 Mana instead of the expected 20x15 (=300). Hence you get 180 less health than expected.

the 18 and 18.7 are attempts to explain this reduction in terms of the stats they're derrived from.
-180 = 18x10sta
-280 = 18.6666666...x15sta rounded down to 18.7 Working your way back it'll introduce a small .5 error: 18.7*15 = 280.5


The above rule is true for all classes.


ReEdit: The above rule is actually already implemented in the Rawr function that calculates health from Sta and mana from intel.
The reason ret and healadin were reducing the 180/280 explicitely is because endoscient didn't use those functions but instead did the conversion calculation himself.
ProtPaladin is in fact doing the calculation itself also, but in a different way which is why I missed it.

---

With endoscient no longer working on rawr or updating the first post... It's probably best to start a new rawr.ret thread ? or can the first post be replaced by the admins ?

It may not actually be an error. I've seen posts indicating that without 2t9, and with certain gear combinations (most often 2t10, bryntoll and TAiaJ, or similar effects that like haste) actually do benefit more from str/haste then str/crit. Unless it is showing it in a rediculous way (very low DPS or massively high dps from either of them) I'd say that it is in fact accurate.

Recent Rawr.ret changes have made the value of haste more prominent.

Haste wasn't being applied to spell GCD's (Cons/Exo), whereas in 2.3.9 it is. This means the 2.3.9 combat simulator manages to squeeze in more abilities over time which it previously didn't do. The change on some added haste (a gem) isn't going to be all that big on dps, but the change from no haste effect on spell GCD (2.3.8) to enabling the effect (2.3.9) has given haste a big boon in the newest version of Rawr.

For single target fights, haste and crit are very close together, and for some gear combinations haste will overtake crit. However seeing how many fights also have adds to deal with, crit tends to be a better stat overall.

I have considered splitting the ret chart into 2. Where now it just has "DPS" following your DPS settings (which would be either single target, or some adds), that would make 'main target DPS' and 'off target dps'.
Main target DPS would be exactly the same as it is now. Off target would simulate damage done to other targets. And you'd have a slider to configure how much % of the time you have extra targets in range for Cons/DS.
Not entirely sure yet if that's a good idea or a really bad one. Ideally, for the Rawr users, the model should be as intuitive/non-confusing as possible. Rawr.ret needs some improvements in that area already, adding a split stat might be more confusing than it's worth.

Something's still up with the haste stat value going crazy, if you load me up in rawr 2.3.9 and equip the T10.264 chest then the value of haste suddenly doubles but if I equip both the T10.264 shoulders and the T10.264 chest the value of haste stays where it should.

Equipping the T10.277 chest doesn't create this issue.

My profile was created by loading one of the BiS.xml's from the bis thread and then reloading my own char from the armory on top of that and rechecking the appropriate buffs, TAiaJ is modeled as per that thread as well.

If you look at the stat graph, you'll likely see that the line for haste isn't straight. For a sim/spreadsheet, haste will change value sharply in steps because at very specific values of haste the number of abilities you can use in a given timeframe changes, or which abilities you can use (since it affects the GCD of Cons and Exo, as well as increasing the proc rate of 2pc T10). While this scenario doesn't really happen in practice due to latency, reaction time, and various fight mechanics, it's hard to overcome it and draw a "straight" line via a sim. You just have to stop taking the value of haste so literally from them and accept it as imperfect.

It's not broken, I believe it will always choose the highest DPS rotation from the rotation list. The stats tab shows you which on that is.

You can force a lower DPS rotation by deleting the others.

Yea I know it will change based on fight lenght, sweet spots and whatnot but something is wrong when it doubles.

Noticed this yesterday, too. It is broken OR we've hit some nirvana with FCFS. I only have 1 rotation in my Rawr right now. The default CS > HoW > Jud > Con > DS > Exo, giving me a simulated 9507dps. When I change it to my real rotation of Jud > DS > CS > HoW > Exo > Cons, I still have 9507. When I change it to Exo > Con > HoW > CS > DS > Jud ... I still have 9507.