Author Topic: Do-it-yourself weapon comparison and ranking (upd Nov22 '04) (Read 93022 times)

Coprolith · « **Reply #15 on:** January 10, 2004, 01:22:38 AM »

I like the idea about the additional line for 2H'ers, Argach, questions about a comparison between a dw combo and a 2HB come up regularly after all.

Lets not get overzealous about adding too many features tho. The complexity of the calculations increases with the number of variables whereas the benefit of adding variables decreases. The most important variables have been resolved: adding proc dps, ATK vs damage, HtH vs 1HB/2HB/1HP and BF. All other variables tweak the calculation by %'s or even fractions of a %. A big issue no one's come up with yet is mob level. I would've liked to put that in, but the issue is simply too complex and the relation between dps and mob level is not known, not to mention ambiguous. You can have different dps against 2 different mobs of the same level.

So any additional features in the pipeline are going to be handled by complexity first and effect second. SCF will be easy. CF will be more complex. Augments have a bigger effect, but until we can figure out a simple and solid way to implement them, you'll have to live with the fact that proc augments get the same resist modifier as the main proc.

At the same time i also have to consider those who are less experienced with dps calculations and using spreadsheets in general. Making them fill out 20 fields before getting a dps number turns the exercise into something like filling out a IRS form.

Right now im concentrating on making the sheet easier to integrate with Yiktiki's. I've already begun writing macro's to do most of the calculations. With macros it'll be easier to integrate and add new features as well.

/hugs

Coprolith · « **Reply #16 on:** January 11, 2004, 10:47:10 AM »

New version, get it here

Changes/fixes in v0.92b

- Changed method of calculations. The calculations now use VB macros. this has no visible effect to the user, but it makes it easier to add new features, and integrate the sheet into Yiktiki's.
You may have to enable macros in Excel's options in order to get this to work (you'll prolly get asked upon start up). No worries, i wont erase your harddisk

. The macros are not protected so feel free to look around.
- Changed the 1Hed DB into 3 separate sheets for 1HB, 1HP and HtH.
- Added another line on the stats and skills page that allows you to directly compare a dw combo with a 2HB (thanks Argach for the suggestion)
- Added SCF
- Added every weapon from Yiktiki's sheets (made a few corrections here and there as well). Also added a few new weapons.

Still in the pipeline:
- Adding augments
- Adding special weapon bonuses such as epic haste and avatar.

/hugs

Hrann · « **Reply #17 on:** January 12, 2004, 01:25:00 PM »

This is great.

I'm not sure if it's me or not, but the proc damage does not seem to be changing when you change the SCF value.

Coprolith · « **Reply #18 on:** January 12, 2004, 02:06:22 PM »

Defined the multiplier for SCF as an Integer. it always got rounded down to 1

Sheet updated

Aggy · « **Reply #19 on:** January 13, 2004, 09:24:36 PM »

Cop: SWEET job, man...very useful! Bewtween you and Yik I think the BL community should be the best informed of ANY of the classes when it comes to eqipment...aprticularly weapons.

You guys ROCK.

Hrann: I like your thoguht process with regards to coming up with the best possible weapon combo for a non-raiding BL, but I don't see how you can really count the procs in your weapon comparisons. Adding two 125-point procs to your weapons is a great way to increase DPS...but the same procs could be added to the raid-dropped weapons, too...

Coprolith · « **Reply #20 on:** January 14, 2004, 10:38:16 AM »

Another small update.

Sempai kindly send me some logfiles of him attacking a red con L70 mob from behind. Based on the results of that parse i've made 2 changes to the calculations that i hope will improve them

- Changed the way proc damage is calculated.
By adding Sempai's data on proc rates to my own i'm getting the very strong impression that the return you get from DEX above 255 diminishes rapidly, to the point where it is no longer possible to even see changes in the proc rate underneath the statistical variance at all. I'd say its even possible that the average proc rate has been hard coded with a cap. I say this because at 255 dex and above the proc rates always average out to neat numbers: 2 procs per min for the MH and 1 proc/min for the OH. The threshold 255 DEX is where the old linear increase in procrate reached those same numbers. That's far too neat for me to believe in coincidence.
I therefore coded proc rates into the sheet in the same way, linear increase up to 255 DEX, constant at 2 procs/min MH and 1 proc/min OH above 255 DEX
- Thanks to Sempai's data, i've been able to at least get a rough approximation that mob level has on hit ratio. I've added this approximation to the calculation, as the effect of mob level on hit ratio makes a considerable difference to melee dps. For more on this, see my latest update of the effects of ATK thread in the library section.

Link to latest version: dps093b.zip

If anyone else has high quality, well controlled dps parses i'd be very grateful if you'd send them to me. After all, its only by comparing the calculations to actual parsed dps that the calculations can be improved.

/hugs

edit: fixed link

Sempai · « **Reply #21 on:** January 14, 2004, 03:30:22 PM »

I have to give credit to Mahes for the mob selection. He is the one that got me the detail I needed to do this test.

I have one minor issue with the spreadsheet.

Some procs cannot have a critical blast, such as the Shinai of the Ancients, even if you have SCF3. (It has a debuff over time component). Therefore, the dps calc is going to be slightly wrong for these types of weapons. Perhaps there is a way to handle this, and I have not found it yet.

For my setup, ED primary and Shinai secondary still seem to be the better dps setup for me as your spreadsheet indicates. I know that most people dislike relying upon procs for reliable DPS, but every analysis I have seen supports this theory. It is this weapon setup that I am working to parse long and hard to prove it once and for all.

Thanks for the great work. Let me know what else I can do to help.

Coprolith · « **Reply #22 on:** January 14, 2004, 04:40:44 PM »

Quote

Some procs cannot have a critical blast, such as the Shinai of the Ancients, even if you have SCF3

Aye, i noticed that. Its just one of those approximations we'll have to live with for now. There are simply too many variables involved in dps to ever get it completely right.
The only work-around I can give for now is that if you are handy enough with a spreadsheet and you know a particular weapon cant crit proc you could manually set the SCFfactor for that weapon to 1.

One day when im completely satisfied with the sheet and we got it integrated with Yiktiki's we can go over the "specials", and manually change them in the DB rather then just copy the formulas. I think that might be a better way to do it, otherwise we'd be adding column after column of input data and the sheet would become very untidy.

/hugs

Hrann · « **Reply #23 on:** January 14, 2004, 09:33:46 PM »

Aggy, re: secondary procs. I've often wondered what sort of dps is added when you have a second proc on a weapon. It seems to me that the dps jump is not as great as if you added a first proc, but I can't say for sure. If it does add that much to a weapon, it would be a nice-to-have in the spreadsheet in the future.

Hrann · « **Reply #24 on:** January 14, 2004, 09:54:20 PM »

I was just playing with the file. It appears that entering the same data on both the first sheet and the sheet with the table, you do not get the same dps answer. For example, put in the numbers from one of the HtH weapons on the StatsnSkills page and you will not get the same MH dps.

Coprolith · « **Reply #25 on:** January 14, 2004, 10:49:29 PM »

Are you sure Hrann cuz im seeing no differences on my comp. Did you update all fields correctly when you changed weapons (the Type field might be easy to miss if you change from one weapon type to another)

If that checks out, could you reproduce the bug for me and send me the details, or better yet, send a copy of the sheet via email?

Thx for testing

/hugs

Hrann · « **Reply #26 on:** January 15, 2004, 03:34:46 PM »

OK, my mind must not have been working particularly well yesterday - I think I hadn't changed the 1HB to HtH on the first sheet when I was inputting a new weapon.

The only minor error I see now is that the LDoN weapon is listed with its old name Slave's Hidden Eitchaka, instead of its proper name, Slave's Hidden Orb.

Kylaz · « **Reply #27 on:** January 15, 2004, 08:32:59 PM »

Damn - wanted to give it a try, but don't have excel at home (EULA respecting beastlord here

) Can anyone post the link for the free excel reader that I think exists ?

Yiktiki · « **Reply #28 on:** January 15, 2004, 11:41:45 PM »

Its in my sig.

Fruud · « **Reply #29 on:** January 16, 2004, 04:18:38 PM »

This is great, I appreciate Coprolith you putting this together and Yiktiki your contribution to the beastlord community is to say the least, well known I also want to thank you.

I hope you two can collaborate and integrate each other’s tools. Good luck.

I only have one small suggestion. Could the most recent version of dps093b.zip for download be placed in the first post? The link in the initial post was dead and I had to go several posts down till I found a working link for the download.

Regardless, great job, kudos’ to you both.