Theck's MATLAB TPS analysis - WotLK/3.x

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Postby Blar » Mon Mar 02, 2009 8:19 am

I thought only the first tick of consecration could only resist and that holy shield could only resist, not miss???
Image
Blar
 
Posts: 43
Joined: Sun Jan 11, 2009 12:02 am

Postby majiben » Mon Mar 02, 2009 8:34 am

You've made a terrible mistake! You assumed a boss mob. They're lvl 82 or lower and thus have a much lower hit cap. You only need 6% hit to be capped against those mobs and most people are close to 157 hit rating (131 for alliance).
Amirya wrote:some bizarre lovechild of Hawking, Einstein, and Theck
User avatar
majiben
Moderator
 
Posts: 6999
Joined: Fri Aug 22, 2008 4:37 pm
Location: Retired

Postby Sarkan-ZdC » Mon Mar 02, 2009 8:38 am

I thought that it was proven now, that conseceration can not be resisted? (Or are we talking about partial resists?)
Sarkan-ZdC
 
Posts: 699
Joined: Tue May 13, 2008 4:00 am

Postby Belarkan » Mon Mar 02, 2009 9:11 am

theckhd wrote:The order is Hit > Strength > AP > SP > Stam. Anything not listed has 0 effect on Holy Shield and Consecration.


I'm a bit surprised I've seen a lot of posts where strength was supposed to be the nr1 stat for threat/dps.

Is the model available ?

Edit:
Forget about it, but is it possible to maintain an up to date on the first page ?
Belarkan
 
Posts: 925
Joined: Mon Aug 13, 2007 3:32 am

Postby theckhd » Mon Mar 02, 2009 10:58 am

Majiben wrote:You've made a terrible mistake! You assumed a boss mob. They're lvl 82 or lower and thus have a much lower hit cap. You only need 6% hit to be capped against those mobs and most people are close to 157 hit rating (131 for alliance).

What, you don't AOE tank boss mobs? ;)

That's a good point. So below 130-ish hit rating, Hit is better, above that STR would win out.


Belarkan wrote:I'm a bit surprised I've seen a lot of posts where strength was supposed to be the nr1 stat for threat/dps.

Is the model available ?

Edit:
Forget about it, but is it possible to maintain an up to date on the first page ?

This thread was split off from another one, so the first post isn't actually mine. If Fridmarr or another mod has the ability to splice and/or move around posts, I'd be glad to write a summary post that links to all of the important findings further on in the thread.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7950
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby Fridmarr » Mon Mar 02, 2009 11:54 am

Sarkan-ZdC wrote:I thought that it was proven now, that conseceration can not be resisted? (Or are we talking about partial resists?)


Opposite, at some point between 2.4 and 3.08 consecrate ticks starting falling off the face of the earth. The current theory is they are resisted but not showing the log.

Basically the testing folks have done is to put on lots of hit gear and consecrate a target dummy a few times and count the ticks. Then take off hit gear and do the same thing. Without the hit gear there seems to be a lot less ticks hitting the dummy.

There's been a few threads here in our Advanced forums about it.
Fridmarr
Global Mod
 
Posts: 9667
Joined: Sun Apr 08, 2007 1:03 am

Postby Fridmarr » Mon Mar 02, 2009 11:58 am

theckhd wrote:This thread was split off from another one, so the first post isn't actually mine. If Fridmarr or another mod has the ability to splice and/or move around posts, I'd be glad to write a summary post that links to all of the important findings further on in the thread.


You are now the first post. I can't change the fact that the Topic is not attributed to you (without SQL and direct DB access), but I think your post being first is sufficient for a summary?
Fridmarr
Global Mod
 
Posts: 9667
Joined: Sun Apr 08, 2007 1:03 am

Postby Dorvan » Mon Mar 02, 2009 12:46 pm

Since questions about Reckoning came up in another thread, I took a look at the code here, and I think the Reckoning calculation isn't correct. I agree that

reck_uptime = (1-(1-0.02.*Reck_points.*(1-player_avoid))^(8/swing_timer)

is the right calculation for Reckoning uptime, but I believe it's factored back into Melee DPS incorrectly. Currently, Reckoning is factored into swing time as follows:

overall_swingtime = weapon_swing_speed*(1-0.5*reck_uptime)

If you simply spot check 0 and 1 this formula gives the right results (overall_swingtime = weapon_swing_speed for reck_uptime = 0, overall_swingtime = weapon_swing_speed*0.5 for reck_uptime = 1), but the values in between don't scale correctly. For example, at 50% reckoning uptime, overall swing speed should be weapon_swing_speed*2/3, but it's being calculated a weapon_swing_speed*.75.

I believe the correct formula should be:

overall_swingtime = weapon_swing_speed/(1+reck_uptime)

The fix the the existing code is just to replace line 200 of stats_recalc.m with:

Reck=1-(1-0.02.*Reck_points.*(1-player_avoid))^(8/swing_timer);

and line 74 of ability_recalc.m with:

Melee_Swing_DPS = Melee_Swing./(weapon_swing_speed./(1+Reck));
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

Postby theckhd » Mon Mar 02, 2009 1:33 pm

Fridmarr wrote:You are now the first post. I can't change the fact that the Topic is not attributed to you (without SQL and direct DB access), but I think your post being first is sufficient for a summary?

Thanks, that will do nicely. I'll edit it as soon as I get time.

Dorvan wrote:I believe the correct formula should be:

overall_swingtime = weapon_swing_speed/(1+reck_uptime)

The fix the the existing code is just to replace line 200 of stats_recalc.m with:

Reck=1-(1-0.02.*Reck_points.*(1-player_avoid))^(8/swing_timer);

and line 74 of ability_recalc.m with:

Melee_Swing_DPS = Melee_Swing./(weapon_swing_speed./(1+Reck));

Thanks, I'll make those corrections. I stole the formula from Jonesy's spreadsheet, and worked it in the way I thought he was using it, but I hadn't had a chance to test it at all yet to see if it made sense.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7950
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Postby Adanel » Mon Mar 02, 2009 2:57 pm

Dorvan wrote:reck_uptime = (1-(1-0.02.*Reck_points.*(1-player_avoid))^(8/swing_timer)

is the right calculation for Reckoning uptime

Assuming 8s duration for reckoning is a dangerous simplification and will make reckoning look ~50% better than it really is with last laugh.
User avatar
Adanel
 
Posts: 46
Joined: Fri Dec 19, 2008 6:07 am
Location: Brazil

Postby Dorvan » Mon Mar 02, 2009 3:02 pm

Adanel wrote:
Dorvan wrote:reck_uptime = (1-(1-0.02.*Reck_points.*(1-player_avoid))^(8/swing_timer)

is the right calculation for Reckoning uptime

Assuming 8s duration for reckoning is a dangerous simplification and will make reckoning look ~50% better than it really is with last laugh.


Good call that the 8 sec duration isn't correct....it couldn't be that easy, could it? :P I don't think we're using any other pseudo-haste effects at the moment, so it should be correct to right it as:

reck_uptime = (1-(1-0.02.*Reck_points.*(1-player_avoid))^(min(8,4*weapon_swing_speed)/swing_timer)

right? Thanks for the catch :)

Of course, this model also doesn't take into consideration any boss actions aside from their standard melee swing, and even then doesn't consider parry haste.
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

Postby Adanel » Mon Mar 02, 2009 3:09 pm

Dorvan wrote:
Adanel wrote:
Dorvan wrote:reck_uptime = (1-(1-0.02.*Reck_points.*(1-player_avoid))^(8/swing_timer)

is the right calculation for Reckoning uptime

Assuming 8s duration for reckoning is a dangerous simplification and will make reckoning look ~50% better than it really is with last laugh.
reck_uptime = (1-(1-0.02.*Reck_points.*(1-player_avoid))^(min(8,4*weapon_swing_speed)/swing_timer)

right? Thanks for the catch :)

It's almost right, just replace 8,4 with 8 (it was a typo probably). (NVM, i got how matlab does it)

The only way to know the real uptime is with simulation due bloodlust and any kinda of haste proc or use here will be calculated as average, but anyway even for broken promisse we're looking for more than 4 swings in a 8s interval raid buffed and the real results won't be much different.
User avatar
Adanel
 
Posts: 46
Joined: Fri Dec 19, 2008 6:07 am
Location: Brazil

Postby steadypal » Mon Mar 02, 2009 3:34 pm

well given the new tier 1 talent is worth a darn.. and we max it out


we can only do 2/3 crusade, is 2/3 crusade and 5/5 conviction better than 5/5 sotp, and 3/5 conviction?
steadypal
 
Posts: 1206
Joined: Sat Mar 15, 2008 12:28 pm

Postby majiben » Mon Mar 02, 2009 3:35 pm

The new T1 talent is hardly worth calling worth a damn.
Amirya wrote:some bizarre lovechild of Hawking, Einstein, and Theck
User avatar
majiben
Moderator
 
Posts: 6999
Joined: Fri Aug 22, 2008 4:37 pm
Location: Retired

Postby Dorvan » Mon Mar 02, 2009 3:44 pm

Like Maj, I strongly disagree with your premise (5/5 Divinity is far from essential), but for those interested here are the numbers:

Code: Select all
Spec                    DPS        TPS
SotP+3/5Convic         3190        8247
5/5Convic+2/3Crusade   3198        8247


No, that's not a typo: for the gear set up included they came out exactly equal on tps, with 2/3 crusade slightly higher for dps. Again though, I really can't see valuing a point in Divinity over a point in Crusade.
Image

WHAT WOULD BEST DESCRIBE YOUR PERSONALITY?
Moonlight Sonata Techno Remix
Scriggle - 85 Fire Mage
Fizzmore - 81 Mut Rogue
Adorania - 80 Disc Priest
User avatar
Dorvan
Maintankadonor
 
Posts: 8462
Joined: Wed Apr 04, 2007 3:28 pm

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: Brokenone, Exabot [Bot], Koatanga and 1 guest

Who is online

In total there are 4 users online :: 3 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 380 on Tue Oct 14, 2008 6:28 pm

Users browsing this forum: Brokenone, Exabot [Bot], Koatanga and 1 guest