Hamlet

[Balance] WrathCalcs

969 posts in this topic

If you back to the beta, I'd looked at this some then. Even at constant haste, you sometimes get variable ticks on your extensions, due to recomputing the remaining number of ticks and rounding. The result will depend on when exactly in the tick cycle the extension occurs. I don't know of any good way to model this, or to incorporate it into playing--it's basically random for practical purposes.

Share this post


Link to post
Share on other sites

It's not though. The most bizarre facet behind this glyph to me was that the number of ticks for both Moonfire and Sunfire were either exactly equal or they were within 1 tick of each other. This was consistent with every single test I ran, except in cases where I had cast 1 more Moonfire/Sunfire than the other.

If I were always casting the DoT then 3 Starfires immediately, I could understand this behavior. However, in practice there is a large gap between Sunfires and Starfires, which you would think would throw off the entire tick-rate. In practice however, that doesn't appear to be the case. Therefore, as long as your haste remains constant I believe the number of additional ticks gained through Moonfire and Sunfire is also constant. As long as you have NG up every time you cast a DoT and every time you extend it (which we do), then the number of ticks gained from the 3 extensions should be the same.

I have no way to back this up yet, but that is how it appears to be working based on the tests I did on a training dummy. It's either that or this is just simply a massive coincidence.

Share this post


Link to post
Share on other sites
It's not though. The most bizarre facet behind this glyph to me was that the number of ticks for both Moonfire and Sunfire were either exactly equal or they were within 1 tick of each other. This was consistent with every single test I ran, except in cases where I had cast 1 more Moonfire/Sunfire than the other.

If I were always casting the DoT then 3 Starfires immediately, I could understand this behavior. However, in practice there is a large gap between Sunfires and Starfires, which you would think would throw off the entire tick-rate. In practice however, that doesn't appear to be the case. Therefore, as long as your haste remains constant I believe the number of additional ticks gained through Moonfire and Sunfire is also constant. As long as you have NG up every time you cast a DoT and every time you extend it (which we do), then the number of ticks gained from the 3 extensions should be the same.

I have no way to back this up yet, but that is how it appears to be working based on the tests I did on a training dummy. It's either that or this is just simply a massive coincidence.

Now that I think about it, I'm not totally sure why I expected it to depend on exactly when the refresh occurred, so long as the tick rate is constant. After 1 refresh, you'll wind up with however many ticks you would have bad on a base 21-second DoT, regardless of the precise timing. The only weird issue is if tick rates change.

Ok, haven't had time today to finish off the multiple items with same item ID issue, uploading the change that puts a blank entry on the reforging dropdowns for now.

Did you ever upload a TreeCalcs with updated item list somewhere?

Share this post


Link to post
Share on other sites
Did you ever upload a TreeCalcs with updated item list somewhere?

I'll get one uploaded with the recent changes shortly (next couple of days). Wanted to make sure OO was working properly before I did all of the stuff for it again.

Share this post


Link to post
Share on other sites
Ok, haven't had time today to finish off the multiple items with same item ID issue, uploading the change that puts a blank entry on the reforging dropdowns for now.

Not sure if this new version wrathcalcs 110124 PTR is intended to be usable on OO but the reforging value cells, the (sub)total value cells of spirit, crit, haste, mastery and hit on top of the sheet and in column C as well as the stat summaries of spirit,crit, haste, mastery and hit show err502. Also there is no value from the table taken if you choose a stat for reforging.

I recommend that all OO users still stick to 110118 PTR till the error is eliminated.

Ready to continue with bug testing on future versions if I dont get banned w/o reason again.

Share this post


Link to post
Share on other sites

Now that I think about this more--what precisely does happen when you extend a DoT?

Say I have a Moonfire with a 1.8s tick time, 18s duration. On cast I get an 18s debuff with 10 scheduled ticks. Somewhere in there I cast SF to extend. Now 3s are supposed to be tacked onto the end of the debuff. Since that's not integral, I think it rounds to 3.6 sec. So I get a 21.6s, 12-tick DoT with one refresh (best I can remember from what I observed on beta).

But what happens on the second refresh? I doubt that it records that the length of the DoT "should" be only 21s and extends to 24. I imagine it tried to add 3s again, which will again be rounded to 3.6. Is this what you were getting at above, ace?

In other words, if your MF tick rate is between 1.2s and 2s, you should always get exactly two ticks from a refresh?

I'll see if I can check this in-game, before I worry about changing haste buffs.

Share this post


Link to post
Share on other sites

More of a 'why does this happen' question than anything else.

When I toggle Bloodlust 'on' my current stats go massively into the negative.

eg

Without BL


Crit rating: 654 1495.33 DPS
Hit rating: 1759 1494.48 DPS
Haste rating: 2037 1495.81 DPS
Spirit: 1463 1494.39 DPS
Intellect: 4547 1497.30 DPS
[/HTML]

With BL

[HTML]Crit rating: 654 -4861.72 DPS
Hit rating: 1759 -4862.57 DPS
Haste rating: 2037 -4861.24 DPS
Spirit: 1463 -4862.66 DPS
Intellect: 4547 -4859.75 DPS
Mastery: 490 -4861.64 DPS

Now, I'm assuming that popping Lust doesn't immediately cause all my gear to become worthless, so I was just wondering if anyone knew what was going on here?

This is with the 110113 PTR version of the sheets.

edit: Thanks Hamlet - I don't think I'd ever hit F9 on WrathCalcs before. Made quite a lot of difference to the numbers I'm seeing!

Share this post


Link to post
Share on other sites

Need to refresh the tables by pressing F9 (Cmd-= on Mac) anytime you want to find the DPS/MP5 values of your stats with a new setup.

Share this post


Link to post
Share on other sites

Thanks, figured out what the problem is (and why it's causing errors in the first place). I'll try to get it fixed tomorrow.

Share this post


Link to post
Share on other sites

With 1813 haste in form, MF tick rate is 1.669s. Unextended MF is 11 ticks. A base 27s DoT would get 16 ticks. Extending a MF with no NG involved gets me 17 ticks. This seems to support what I was saying above.

When the MF procs NG and I extend immediately, I can get up to 20 ticks, but there seems to be some variation.

edit: I actually just got 18 with no haste changes involved.

Share this post


Link to post
Share on other sites

I'll be honest Hamlet, I have no idea what's going on. I'm really at a loss to explain the method behind the madness that is the extensions, but there is definitely a method to it. My best guess is that it takes the remaining duration, adds 3 seconds to it, then recomputes the tick rate based on your current value of haste. For example, my tick rate (without NG) is 1.66 and let's say there are 13.66 seconds remaining on the duration of the spell when Starfire lands to extend it. 13.66 seconds + 3 would be 16.66 and with a tick rate of 1.66 seconds, it comes to 10 ticks when there were 9 ticks remaining on the spell. I believe there is some amount of rounding involved though. So say we were to extend the spell when there are 14.774 seconds of duration left, which is normally 9 ticks left. 14.774 + 3 seconds would be a duration of 17.774 and with a tick rate of 1.66 you would get 10.7 ticks, rounded to the nearest number of ticks, which would make it 11. That is a gain of 2 ticks, with no change in Haste. It is simply a change in the timing of the extension.

It definitely does not always add a constant amount, as I have observed getting additional ticks that were not multiples of 3 (I remember one specific example being +5 ticks for example). This is the only way I can explain the effects right now. I do believe it comes out to an average though, otherwise I wouldn't have such a clean split between Moonfire and Sunfire. I simply don't know how to express that average at the moment.

Edit:

It just dawned on me that if what I suggested above is true, you might be able to express the average number of ticks based on the cycle length of our rotation since we will always end up clipping the DoT just before Eclipse ends. For example, if it takes you 21 seconds to go from Pre-Lunar to Pre-Solar, and you're tick rate is 1.66 seconds, you will end up getting on average about 12.65 ticks, correct? This assumes that the tick rate is constant throughout the entire rotation and that there is never any downtime of Moonfire or Sunfire, which there isn't if we are always clipping. There is an issue of course with Haste procs throwing that off, but I can't see how this would be any different from how things in WrathCalcs are now.

Share this post


Link to post
Share on other sites

Tweaking GoSF a little, although it didn't really change all that much. If you select GoSF and Moonfire on each Eclipse, then it assumes all Moonfire damage is affected by Eclipse and NG, and 100% uptime with 2 casts per cycle.

WrathCalcs 110126 PTR.xls

Share this post


Link to post
Share on other sites

Update for the day. Findings about Euphoria in the other thread included (doesn't really change much). I also took the chance to cut some unnecessary rotation stuff, like DoT/Starsurge options that are no longer relevant. Might help make more of the rotation page comprehensible to people besides me :/ . Tried to add in the 4T11 effect on DoT's as well. It's really optimistic right now though. Things look really nice if you assume MF/SF is up 100% of the time, always with Eclipse and NG, and benefits from 100% crit for a few seconds each cycle. I'm just not sure how meaningful it is in practice.

WrathCalcs 110127 PTR.xls

Share this post


Link to post
Share on other sites
edit: I actually just got 18 with no haste changes involved.

If you saved a log, you may find that one of your SF landed very shortly before an MF tick, and server bugs gave you an extra tick

1) SF lands with 5 MF ticks remaining. Server checks to see if that should be extended to 7.

2) MF ticks

3) Server figures out that yes, MF ticks remaining should be extended to 7.

Magi discuss a similar behavior for ignite munching:

The time between the actual crit and the ignite applying is the server verifying that, yes, your character has the talent necessary before actually applying the debuff. Proc-based talents from all classes suffer this ~0.5s delay. This delay isn't a huge problem for other class-based procs, but strange things with ignite happen with certain sequences of events:

[00:13.872] Spell Crit on Target

[00:14.381] Ignite Tick

[00:14.461] Ignite refreshed on Target

In this situation, the Ignite Tick is a 'free' tick - since it falls between a valid crit refreshing the ignite ... and the actual refresh event

I haven't noticed a log that shows this for GoSF (which I'm not using). In 3.x I'd commonly see something similar for Lifebloom (too many total ticks from a slow-stack).

Share this post


Link to post
Share on other sites

Apologies for the lack of updates this week guys, drowning in work at the moment. I'll try to get the OpenOffice bugs sorted out for the last update shortly.

[EDIT]

Quick fix for the profiles as the rotation section changed:

WrathCalcs 110201 PTR.xls

Share this post


Link to post
Share on other sites

I don't I think I had anything else that I was going to do at the moment, you can update things whenever. Assuming the patch is coming on Tuesday, that's probably when I'm going to put a new main version on the guide thread on this forum.

Share this post


Link to post
Share on other sites

Right, I'll do my best to get everything squared away before the patch, then.

How is the profile feature working for those of you who can use it?

Share this post


Link to post
Share on other sites

Profiles seem to work fine for me. Copying my profile from an old sheet to a new one is slightly non-obvious. I think I had to

0) Quick look to make sure the profile format hasn't changed (it is bound to happen sooner or later).

1) Make a new profile on the new sheet.

2) Copy/paste the profile from the old sheet, over the new profile.

An export/import for profiles could probably take care of all of that. I'm not VBA-savvy enough to want to try that.

Share this post


Link to post
Share on other sites

Yeah, people ask about that a lot. If there's a simple way to add a user-friendly import/export that would be cool. Otherwise can at least post instructions for people.

Share this post


Link to post
Share on other sites

FoN against a Raider's Dummy in SW

Naked, MotW, Moonkin form, 133 FoN swings (2.5 casts, I reset Recount once while FoN was up)

Hit 34%

Glance 26%

Parry 17%

Miss 10%

Dodge 8%

Hit (Blocked) 5%

Crit 1%

Dressed with 14.02% spell hit (all from Spirit), 0.0% Melee Hit, 0.0% Expertise. 110 FoN swings

Hit 56%

Glance 26%

Dodge 3%

Parry 7%

Hit (Blocked) 6%

Crit 3%

I did another cast at 14.02% spell hit (50 more swings), and finally got 1 miss.

It seems likely that spell hit helps treants avoid misses, and parries. It might help reduce dodges (needs more testing). Probably doesn't help Glance or Block. That second set had 9% crit rating. Inconclusive on how much it helped the treants.

Share this post


Link to post
Share on other sites

I've been wondering that myself. The weird thing is that even at spell hit cap, there's still the occasional miss. Dopefish never posted his actual data, I wonder if he saw the rare miss at 17% and assumed that spell hit was doing nothing.

At any rate, I'd been pretty convinced on spell hit anyway, but this still makes it unclear what numbers to put into the sheet for scaling. I'll could just change it to use spell hit for now (but this would make miss/dodge rate 0%).

Share this post


Link to post
Share on other sites

Were you sure to place Treants at the same spot every time and/or they moved to the same location when they started attacking? If memory serves, bosses shouldn't be able to block or parry from behind (but they can still dodge).

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.