Negative EP Values

[GoogleCodeExporter]

What steps will reproduce the problem?
1.  Certain templates/priorities together will generate negative values in 
EP tests.  This is impossible.  No stat or set bonus can hurt your DPS.  
It may generate 0 benefit, but that implies a 0 stat weight.

What is the expected output? What do you see instead?
The worst a stat should be able to be rated is 0.  This implies 0 
improvement.  Since no stat can actually make your DPS go down (only go up 
very slightly or not change), a negative value isn't sane.

What version of the product are you using? On what operating system?
Version 1.0 on Vista x64

Please provide any additional information below.
I first noticed this issue in 0.9.9, where it was rating haste rating and 
the 4PT7 bonus as negative values.  In 1.0 haste is now rated as a 
positive value (fine), but 4PT7 is still negative.  I could understand it 
being of 0 benefit potentially because the RP generated by the bonus could 
be wasted, but it should NEVER be negative.  It can't HURT to have too 
much RP, it just is wasted and doesn't give a benefit.

Attached are the files I used.  Lysera.xml is the character file.  Lysera-
DW is the template.  The priority I used was Frost.xml.

Original issue reported on code.google.com by MCla...@gmail.com on 20 Aug 2009 at 1:14

Attachments:

• Lysera.xml
• Lysera-DW.xml

[GoogleCodeExporter]

Forgot to mention that the other values during my EP test were as follows:
Presence: Blood
Sigil:  Awareness
MH: FC
OH: RI

100 Hour Test with a latency of 150ms.  All buffs except Draenei.  In 0.9.9 I 
had 
tried to do this with a 5000 hour and the results were identical.  I haven't 
had the 
time to attempt a test of that length in 1.0 yet, but the fact negative values 
are 
being displayed at all is a problem.

Original comment by MCla...@gmail.com on 20 Aug 2009 at 1:20

[GoogleCodeExporter]

I will take a look.

Original comment by Fabien.Arpinon@gmail.com on 20 Aug 2009 at 7:52

• Changed state: Accepted

[GoogleCodeExporter]

Without looking if there is a problem in the code, maybe it's because 4PT8 give 
you
more runic power, so more Killing machine proc eaten by FS, so less Rime used 
with
the frost priority. I will try to enable the report display during EP 
calculation to
see if it's this.

Original comment by Fabien.Arpinon@gmail.com on 20 Aug 2009 at 9:29

[GoogleCodeExporter]

I thought the same thing at first, however, that shouldn't be an issue though 
due to the priority layout of 
Frost.xml.

Frost Fever
Blood Plague
KMRime
Obliterate
Blood Strike
Frost Strike
Rime

The only time it should be attempting a Frost Strike is if there is, 
essentially, no other move to be made.  It 
shouldn't be concerned with MaxRP or not, only worrying about if there are 
runes up.  This means that the 
excess runic power shouldn't be an issue; at worst it will be wasted because 
nothing will use it.  The only 
thing it is prioritizing over is Rime, which is the correct things to do as 
you're supposed to wait for a KM proc 
to use it (at which point it is prioritized differently).

Original comment by MCla...@gmail.com on 20 Aug 2009 at 10:28