New version of my universal script

[Leafkiller]

I just created a new version of my universal script
("universal.script") and checked it into Mew, along with the test
profile I used for the optimizations (I got the stats by creating a
toon on the Ptr).

The change log I posted with the check in:
[code]New version of my universal script. It combines both the most
current (6/23)
default Mew script and an newly optimized version of my script (based
on the
default script, my 4.1 script and some additional optimizations). All
the
optimizations were done with a single Mew profile - 378Leafhitexp.mew
which I am
also checking in. The new script takes advantage of params to select
with script
to run - it is documented at the top of the script.[/code]

Usage instructions (documented in the script):
[code]// Leafkiller universal script
// Usage: In the "Custom Strategy Arguments:" box enter the following
// [ ] default script (empty)
// [atramedes ] default script, atramedes fight
// [leaf ] my script
// [leaf atramedes ] my script, atramedes fight
// [exp ] experimental script (same as leaf currently)
// [exp atramedes ] experimental script, atramedes fight (same
as leaf atramedes currently)[/code]

Using the 378Leafhitexp.mew profile I am seeing about a 270dps
increase over the default script on a 5 minute patchwork fight. A
little more with the Atramedes options.

[Yawning Angel]

On Fri, Jun 24, 2011 at 1:30 AM, Leafkiller <roy.d....@gmail.com> wrote:
> Using the 378Leafhitexp.mew profile I am seeing about a 270dps
> increase over the default script on a 5 minute patchwork fight. A
> little more with the Atramedes options.

I just went and stole all the good bits, and now the scripts do near
identical (as in, the difference is not statistically significant)
DPS. A surprisingly large amount of the gains were from using Heroism
right at the start of a encounter. Not sure what raids actually do
these days, and behavior should probably be configurable, though
changing the script to switch from early to late is 2 bytes so I'm not
all that worried about it.

Unless you have a strong objection, it's probably a good idea to nuke
your old script(s) from the samples directory.

Regards,

Yawning

[Leafkiller]

I tried to nuke one of the old scripts already (universal1), I guess I
just did it wrong as it is nuked from my machine but not from the
repository. Can you dump that one, since I don't have it available
anymore? Also, would you dump the atramedes script? Clearly I am not
using Tortoise correctly. I should really redo my environment to use
subversive and stop being lazy...

Some thoughts on what I saw:

Heroism is worth more at the start because it is overlapping with
Berserk. That ended up in my code because of the Atramedes work I did
and I did not think about it much last night when merging the scripts.
If we optimized the Berserk code we could move Heroism back to the end
of the fight. We would simply need to look at fight length, not unlike
players try to do.

Energy pooling was pretty significant in my testing (I used 70 energy
in my 4.1 script). I tested with a profile which was hit/exp capped
but had very low haste and the dps gain was substantial. We have not
spent that much time discussing the interrelationship between the
rotation and the toon's stats, with most people treating the rotation
as fixed and the stats as the variable (as they search for the
"mythical" best secondary stat). It looks like 4.2 wants a higher base
energy pool than 4.1 did. This can be accomplished by letting it get
higher and capping hit/exp to help avoid losing energy due to misses,
or by having higher haste.

Removing the 6 second Rip guard time from SR refreshing reflects the
buff to SR compared to Rip. While Rip still has higher dpe, the 6
second window traded off a lot of SR uptime in order to minimize Rip
downtime. With the relative values shifting, that tradeoff is now a
dps loss. This may also be a reason higher energy pooling is a dps up
- less time to get the Rip back up after a SR.

With FBs back in the rotation and more aggressive SRs, combo points
once again have a premium value to us. It will be interesting to see
how different reforging profiles stack up the the changes in the
rotation.

Did you consider adding the atramedes option to the default script?
Over time we could add several options to it to drive different fight
scenarios. That is about the only difference between our scripts now.
Did I mention that I have to update my script now to include the
latest version of yours? That is cruel and unusual punishment imo.

[Leafkiller]

Never mind on the deleting scripts...I forgot to commit the delete
after I made it - so I am doing it correctly. Using the 378 hit/exp
profile I checked in last night, the default script is now about 7 dps
ahead of my script. If we fold the atramedes code into the default
script, other than the convenience of having an experimental script to
swap between while testing changes, I don't see much value in
maintaining the universal script anymore. There was a time where the
rotations had some interesting differences, but that time has passed :P