Tecton, can you re-upload the last thing you added with the macros put back in? I'm pretty sure I can load/save while preserving the macros, I just can't run them. So they can be left in the sheet for others.

Also, can you add in Lightweave enchant and Leatherworking enchants (don't worry about what values to give Lightweave, that's what I want to work on)? If there are any trinkets or other things you skipped because they had procs that couldn't be handled properly, can you add those in too? Basically, you can see how I handled Power Torrent and Shard of Woe in the version I uploaded before--those should be good methods of handling any similar procs.

Note to self: fix Shard of Woe data table in next version (don't want to change anything locally now to avoid branching). Erdluf suggestions sound good for custom stats. Ace I think has some ideas for compactifying the front page aesthetically and cutting down on some of the blank space--I'll see when might be a good time to try working those in.

In general, how do people think professions should be handled? More dropdowns on the front page like race (which would then enable/disable profession-specific bonuses whereever they are found)? Or just have all profession benefits available and trust people to select only the ones corresponding to what they have?

Finally, there needs to be a way to select Engineering tinkers, which are now separate from enchantments. If I recall, the only slot having a tinker that matters for anything is Gloves (boot/belt/cloak tinkers have no stat or DPS bonuses now), so maybe just add another row to the Gloves box only. Tecton, if you could also throw that in and options for the relevant tinkers (Synapse Springs, Z50 Mana Gulper, Tazik Shocker), that would be cool.

(Also, if you don't have time to do those now, just tell me so I can work on other things myself. Basically using an informal checkout scheme so I don't have to merge things later).

Ok:

• Macros are back in, you can update items from MMO-C's database now. Procs and sets should be added by hand, though.
• Lightweave is in, I could only find one leatherworking enchant that was applicable (the intellect one). If I've missed any, let me know.
• Engineering tinkers can now be added to gloves, I just added a row as suggested. The only one with a value at the moment is Synapse Springs, I wasn't sure how you intended to handle the others.
• Shard of Woe is back on the gear list. I *think* I have all of the trinkets/etc now, but if anyone spots one of value that's missing, shout up.

As far as procs, I didn't go through the trinkets and add those in since I'm not sure what the ICD is now. I remember reading some rogues were seeing 75 second ICDs on their trinkets, didn't want to rush ahead with those before I was sure what the situation was. Do you want to average out haste/mastery/etc procs and on-use effects in the same way?

For trinkets like Gale of Shadows - Item - World of Warcraft I've just put their values in as if they were fully stacked.

I see you added a "Temp Int" column to the front page. Might be unnecessary, but won't worry about it for now, since there might be aesthetic cleanup of the front page later to get rid of some of the white space. Is a bit easier to work with for the moment.

As far as trinkets that merely give averaged-in procs that I'm going to treat as constant stats anyway, that depends on how I go about dealing with the flow of stats. Here's what happens right now (for Int, which is the most complex one):

1) Front page module adds up racial base stat and Int from gear.
2) Old Main page takes (1) and multiplies in permanent buffs that show up on the character sheet (e.g. HotW). This produces an Int value that was used as the input in the old sheet.
3) Basic calcs page takes (2) and divides out everything that was added in (2) to produce TrueInt, which is of course equal to (1). The 1 extra Int for data table computation also gets added here.
4) Basic calcs page takes (3) and adds in all buffs to produce final Int without procs (used only for mana pool size).
5) Basic calcs page takes (3), adds average value of any procs, and then adds all buffs to produce final average Int (used for all other purposes).

The back-and-forth in steps 1/2/3 appeared for now just because the old sheet started with 2, and I plugged Tecton's stuff into that. The issue surrounds the option for manual inputs though. Without manual inputs, steps 2 and 3 could be removed entirely.

If we want manual inputs, what value should they be?
--The same unbuffed character sheet values as before, in which case I basically maintain the current setup and add an option to override at step 2.
--Completely unbuffed values, in which case I can delete 2 and 3 again, but people using them have to have the wherewithal to use values without HotW and Astral Leather included.

Once we decide that, we decide where to include the mean value of procs (imagine a spellpower proc for now, since there's already a separate system for Int procs due to the mana issue).
--Just include them as a constant stat, so they get automatically included in step 1. Simplest, but manual inputs are more confusing now since they need to include the mean value of procs (basically the old system, which was not very user-friendly).
--Insert them in step 3, which is clearer, but takes new machinery.
--Add new TempSP/TempHaste/etc. stats, like the TempInt one, just for this purpose. Probably more added complexity than it's worth.


Trinkets with effects too complex to model simply as a mean stat value get the Shard of Woe treatment--dedicated T/F box in the main page, that reads from the item list.

As far as manual inputs are concerned...

If you're going to maintain the option for manual inputs, it would be best to leave it as it is currently: the unbuffed value that appears on your character sheet in game. That's simply the easiest value from a user's perspective. I personally wouldn't want to break down my "Character Sheet" Int into "True" Int every time I feel the need to use a manual value.

That's what I figured. The real issue is how to deal with temporary buffs though. If the idea to continue displaying charsheet Int somewhere to serve as a check/override point, procs have to be added in after that, which means they can't be modeled as plain stats.

Another option is only have manual inputs in the form of +/- adjustments on the front page like Erdluf mentioned. The idea here is that only advanced users or theorycraft testers are likely to need to drop in manual stat values, and we have the wherewithal to just copy the sheet and manually overwrite stuff if they really need to (which is generally what I do even now if I'm checking spell coefficients or something).

------

A few thoughts on actual rotation modeling. Since the DoT's are practically always refreshed in tandem in a stationary rotation, is it better to just use an NG uptime of 0.5 for each DoT? Alternatively, since IS should really always be refreshed first (note that MF is stronger than IS under both Eclipses), use 0 for IS and 1 for MF?

The only way to get both DoT's NGed at once is if you refresh both at the very end of one Eclipse to proc NG, then refresh both right before the next Eclipse while still under the prior NG. This clips not only the DoT's but the NG as well, so I doubt there's a viable DPS technique here.

As I've mentioned before, I really want to ditch the huge current system for computing NG uptime (which currently runs the whole rotation twice, once just to compute NG uptime and once to compute final DPS). In the case of DoT's, using the constant values mentioned above might be not only a good approximation, but might even be more accurate than what's in there now. I think coming up with a good approximation for the effect on NG on other spells might not be too bad.

I think the +/- adjustments would be great. Honestly, that would make comparing gear differences much easier than it is currently, which is one of the shortcomings that WrathCalcs has in comparison to tools like Rawr. Such adjustments would make manual character sheet values completely unnecessary.

As far as Nature's Grace is concerned...
It took me a minute to think about all that, but it finally settled in (been away from WoW for a while) and I agree that using a NG uptime of 1 for MF and 0 for IS is probably the best way to model it. The only way I can see IS ever getting buffed by NG is if we somehow manage to get our rotations time to the point that NG is always up. That seems very unlikely though.

Edit:
Since Hamlet mentioned some of my aesthetic improvements to the sheet, I figured I'd provide a link to what I was working on. Everything on the right side (after column I) are just notes and leftover scraps of the initial Front End. There are a few things I'm thinking about placement wise, but this should give you a pretty good idea of what I had in mind. Not everything works at the moment, but I think the list I have for what would be needed is accurate.

Here's the sheet.
Thoughts?

I think you can avoid the cyclic NG computation. A full Eclipse cycle generates 200 solar energy and 200 lunar energy (plus a bit of wasted energy at either end, which I think the spreadsheet ignores).

If we can assume that each half-cycle takes at least 15s and no more than 60s, then we can safely say that we get 15s of NG during each half-cycle.

We care very little how much of that 15s was during Eclipse, but I'm going to assume that exactly half of it was during Eclipse, because it makes later calculations easier. If you adjust that number, I'm assuming the impact on overall DPS will be small.

For just the SF part of the rotation (Lunar and pre-Solar) it should be straightforward to compute (in this order):

1) Damage and Energy from 7.5s of Lunar+NG = D1 and E1
2) Damage and Energy from 7.5s of NG without Lunar = D2 and E2
3) Determine time and damage associated with Lunar (no NG) to generate 100-E1 energy. T3 and D3
4) Determine time and damage associated with no lunar and no NG to generate 100-E2 energy. T4 and D4.

Note that the order these occur is actually 3-1-2-4.

For this half-cycle, total time is 15+T3+T4. Total damage is D1+D2+D3+D4.
You'd do the same thing for the other half cycle.

Personally I'd assume that DoTs are not in lock step (they might be for patchwerk, but most fights look a bit different). If the MF strategy is "always refresh", I'd say that you average 7.5/18 MF casts during Lunar+NG. If someone's play style keeps their DoTs in lock step, they might want that as an option (in which case sub-phase (1) always includes an MF, and sub-phase (3) always ends with an IS.

Edit: Note that this never uses a spell's mean cast time. It either uses its NG cast time, or its non-NG cast time. The averages will fall out once you know the total rotation length, but you don't really need them.

I think your computation works fine when the only spells in play are SF and W. But in that case, the "4.5 second rule" would work perfectly anyway. The problem comes in when you start having other spells that are used on cooldown, you need to know what % of those are affected by NG. Note that their casts/cycle don't remain constant with cycle length--they decrease, because cooldowns aren't affected by haste (or in the context of your algorithm--once other spells are involved, you no longer cast a full 15s of NG's Starfire in one half-cycle). As I put in my edit above, just now, I think that's why it's okay that the difference between NG and non-NG cycle length is more than 4.5s when spells other than SF and W are on. And the fact that it's exactly 4.5s in the sheet with all other spells off is a very good check.

What I'm planning to do is:
1) Compute cycle length without NG (this is much simpler than the current first iteration). Call this L.
2) Set NG uptime equal to X = 30/(L-4.5).
3) Compute Y from X based on above equation.
4) Compute real cycle using NG rate Y for all spells.

-------

DoT's:
Even if DoT's are not in lockstep, the first you cast after Eclipse is not affected by NG, and the second is. And those two DoT's are always going to be one of each (in any generally normal situation modeled by the sheet). So the only effect of having your DoT's out of phase is that sometimes MF gets and Eclipse instead of IS. It still seems that MF should use NG uptime P and IS should use (1-P).

This is ignores the fact that sometimes you recast at odd times due to interruption or something. But the current system and the one you suggested (which both use some computed mean NG uptime for DoT's) ignore the fact that the first DoT isn't affected by NG, which I think is more significant in practice.

Although I'm looking at T1 epics from the beta, where my cycle length is 41 seconds. So 1 set of DoT's each Eclipse is basically what winds up happening. At lower haste levels you're going to cast more non-NG, non-Eclipse DoT's. Have to account for that somehow.

For most spells used on-cooldown this is easy. If a sub-cycle has a GCD of 1.1s, the fraction of time spent casting FoN in that sub-cycle is 1.1/180.

You have to do some fudging for for time spent on Starfall and SS, but the impact should be minor.

Hmm, maybe. I wound up spending way too much time on this today, so I'll probably look at that next time I try redoing the model again. I'm not totally convinced it avoids similar circularities to what I run into doing it the current way (remember for example the effect of DoT tick rate on Shooting Stars).

Anyway, here's what I came up with today (attached). I redid some of the NG model and some stuff about DoT's. Other than that I just added the manual stat adjustments on the first page (and the previous stat inputs are more or less unchanged, so you if you want to test things you can just overwrite them and it will work as before). I rearranged page 2 a little as well.

For trinket procs, I think it's fine to just add them as mean spellpower (or whatever) on the item itself. The only downside is that the page 2 stat values won't exactly match armory, but I think that's fine. Anyone who's going to be doing much with those values will understand what's going on--assuming the whole item input system works correctly, basic users won't have to pay too much attention to the listed stat values anyway. It's worth creating a whole bunch of new stats just to handle them. Int is the only exception due to the mana issue, but even then I don't think it needs the full stat treatment with the column on the front page and in the gear table. Can probably just add it up somewhere with conditionals for each item that can has it (in fact I'll just do that right now).

Like I said above, I should take a break from this for tomorrow, so I'm uploading for now in case anyone else wants to work on stuff. Things I can think of that would be handy:
--Fill in proc values for trinkets. If you don't know the ICD, just guess based on WLK principles, I can adjust later once we know.
--If there are any trinkets with more complex effects than that, note them here so I can add special treatment. I haven't looked at the loot lists in detail yet.
--I think working on the aesthetics of the front page would be nice. Aceofsween's design is pretty neat, although it doesn't actually save a huge amount of space due to all the repeated headings. Also, it requires organizing the stats in a totally different way (2 boxes with labels for the two secondary stats on each item, instead of lots of columns with most of them blank). I'm not sure how easy that is--the person to do it would be Tecton, not me, since it's really about hooking everything up again in a different way. If there's no easy way to change, don't worry about it. Maybe some of the ideas can somehow be used to compact things somewhat though.
--Trying to make the page 1 color scheme match page 2 better would be good. In particular, using borders/colors more to clearly distinguish inputs from everything besides inputs.
--Should all dropdowns have a "none" field? Right now you can just delete a field to unequip it, but that might not be obvious to everyone.
--I think two additional dropdowns on the front page for professions would be good. It shouldn't take all that much work to make certain things available/unavailable based on which ones you have selected.
--If you're ever in the mood for what I'm imagining would be a larger new feature, Armory import would be a cool thing to have at some point.

You can put in any or all of those whenever. I'll post again when I want to edit things more myself. This seems to be coming along pretty well.

Quick list of trinkets that I think would be slightly trickier to model than just averaging out an int proc:

Heart of Ignacious - Item - World of Warcraft. The haste on-use removes your stack of spell power. For now, I've just got it as the value of the stacked spellpower, I'm not sure how you'd want to handle the on-use.

Darkmoon Card: Volcano - Item - World of Warcraft. I'm not sure how you'd want to handle the fire damage proc, or direct damage procs/on-use effects in general.

Tyrande's Favorite Doll - Item - World of Warcraft. I'm not sure if the damage is on the GCD, but the mana gain might be something to add anyway.

Sorrowsong - Item - World of Warcraft. 50% uptime spellpower proc under 35% target HP.

[edit] Currently looking at changing the layout to something similar to aceofsween's, although I do think the repeated titles/etc make it look a bit busy, so I'll likely work around that a little. I'll see what I can put together. [/edit]

[edit 2] Ok, so I should have this functional today, but it's taking a good bit of rejigging. If you can hold off making any changes until I've uploaded another copy as I don't really want to have to port this over to another version if I can avoid it. Cheers! [/edit]

Are you just saying I forgot to link some things to the stat value which has the manual adjustment added in? That would be easy to fix.

e: yeah, I haven't changed anything over here. I'll wait until you post something.

I'm hoping to add a custom model for any trinket that requires it, like I did for Reign in the past. Most aren't too bad.

Ok, I've swapped things around (the old stat list is still in the sheet, just hidden, as it's easier to work from something like that rather than having a million SUMIFs all over the place). The stuff that's hidden does need to be tidied up, but I think this should be fully functional and I didn't want to delay it too much.

Still on the short to medium term to-do list, just FYI:

• "None" field on dropdowns.
• Add conditional formatting for gems/etc, (although I'm concerned this might make the page look too busy). I'm not a hugely visual person as you can probably tell from the colour schemes I went with at first, so feedback here is welcome.
• Match the colour scheme better to page 2.
• I will be adding armory import as soon as the new armory is up and running properly.
• Trinket proc values/etc (could use some assistance with this).

Long term, if there is any demand for it, I'd like to look at some kind of automated gemming/reforging calculator and upgrade suggestions.

If I've missed anything with the new format, or anything, just let me know. I'll wait to make any more changes now until you post up you're done with it, Hamlet.

Beta ended today. I did check that the new frontpage module gives exactly the right stat values after gems/reforging for the setup I had on beta.

I never did bother getting numbers on Force of Nature though >:-| . Does anyone have any idea how much damage is done the Treants now based on spellpower?

Hmm, another tricky issue with adding temporary buffs as average normal stats is that it throws off haste breakpoints.

EDIT: wait, no it doesn't.

Spellpower is never displayed as a stat on the front page. Is that important? It only appears on weapons and (now) trinkets and can never be reforged.

Edit:
After looking at the layout Techno has, I see where you're coming from now Hamlet. It's probably a little redundant to have them both in there that way.

Also, the new layout looks good Techno. The color scheme update will probably make it look much better. A few suggestions on that note: is color-coding the sockets by their socket color possible? And on a similar note, could you make it so that the box for the reforged stat on the item (listed with the other stats) is differentiated from the other stats? It's not a big deal, but it might make things a bit easier to differentiate when someone thinks, "wait a second, this doesn't have spirit on it..."

Btw, Box AF:3 says "checklist" when it should probably say "Hit Capped" or something to that nature. Also, I would change that to just give the Hit Rate (in percentage) and then +XXX Hit needed to cap in parenthesis next to it, So it would read something like Hit Rate: 16.40% (367.255) instead of the Yes/No by XXX rating that it currently says. Also while I'm thinking about it, the check for that box is directed to I3 instead of I5, but the calculation of the rating seems to be correct.

Looks good Hamlet, but would apply the same color coding for the sockets to the gem selection and their explanation as well (makes things look a little more polished).

Fixing stuff now:

• I'm now displaying Spellpower in the same cell as Intellect for weapons & trinkets. I didn't think it warranted adding another row.
• I've added a single row between each item, 10 pixels, good suggestion. It does cramp things in the hidden values a little, but we can expand those when we need to check them. I readjusted the guide formatting a little to allow for this.
• MATCH() changed to an array formula, which should remove the crit/hit item issue and is a bit of a better solution anyway, as you say.
• Reforges should no longer display at all if they are invalid.
• !VALUE errors fixed, it was a misplaced bracket when adding the reforge values (i.e. it was adding it outside of the error checking).

I'm trying to figure out some way of preventing reforging of trinket procs/uses, if I can't get it shortly I'll upload these changes. Are the asterisks you added beside trinkets purely for a visual check?

Regarding adding conditional formatting for each kind of gem, Excel by default has a cap of 3 conditions. You can get around it, but it's a bit of work to do so (especially while avoiding macros/etc). I can take it a look at it further down the road.

Didn't get the proc/etc stuff off of the reforge options yet, bear with me on that one. Above stuff is fixed, reforge menus are tidied up.

Latest version:

If the insane requirement for the Chaotic meta remains, could it be worth it using the wotlk-meta? I can imagine the ~20 crit rating being made up by being able to have only 2 blues rest red.

The WotLK meta was updated to the same requirement, from memory.

[edit] Ran out of time for today, I'll add professions after your next draft. [/edit]

You can keep it for today if you want, I don't think I particularly have anything I was going to do.