EQMac.com

For an end-game ranger, which bow is best?

I've heard that most rangers on the server use Tynan's spreadsheet to determine this (see viewtopic.php?t=5810). However, the mechanics of EQLive in 2003 and EQMac in 2012 may be different. I've decided to determine if the values in Tynan's spreadsheet are actually congruent with Al'Kabor instead of just assuming that it is fact. In the tradition of giving back to the community, I'm going to share my results with everyone.

Parse information:
- Each trial is 24+ hours of autofire
- Shooting at Bella Helsin in Katta
- Buffless
- Max AA
- Using Blessed Champion Arrows
- Using Fleeting Quiver
- Worn +170 attack (+180 attack with Stonewood)
- OSX client
- Using iKey autofire
- Parsed with YALP

Sources of error:
1) Bella Helsin is a green-con mob to a 65. This will inflate the reported DPS.
2) Along the same lines, Bella Helsin will never resist a DD proc. This will inflate the reported DPS of the Fire, Air, and TZ bows.
2) I filtered out head shots, so for the purposes of this parse they become a miss instead of a hit. This will deflate the reported DPS.
3) I did not shoot long enough for the results to be statistically significant.

Math:
- How long must I shoot to obtain a standard degree of statistical certainty? Not sure how to calculate this.
- After doing two trials of Air, I got 131.76 and 131.26, so the figures can be considered (at the minimum) +- 1 from the actual values.

Results, sorted from best to worst:
1) Air - 131.5 DPS
2) TZ - 128 DPS
3) Earth - 127 DPS
4) Water - 120 DPS
5) Fire - 116 DPS

Conclusion:
- This is not the "end-all be-all" parse; I'm simply providing this data for your reference. Each bow may perform differently on a dark blue con XP mob, or a red con raid mob.
- The order is likely accurate since the DPS differential between bows is 3 or more (with the exception of the TZ/Earth bows, which should probably be considered approximately equivalent).
- My ranking is congruent with Tynan's ranking with the exception of the earth and water bows being reversed. I feel that this discovery in particular makes the parse worthwhile, since the earth bow seems to be easier to obtain than the water/fire.

Cheers.

Good stuff. I wonder, though, whether haste rounding changes the picture at all. My (poorly informed) suspicion is that at 100% haste (i.e. raid buffs without Riz) the earth bow will gain a little bit of DPS relative to the other four, and that at 125% haste (riz) the air bow will pull even further ahead.

By the way, where on earth did you get the air and fire bows? Those can't be easy to come across even for a guy with your considerable resources.

_________________
Ryoku Wing, Swift Tail of Temerity

Soft-skinned alt scum:
[ Ravenwing | Syovan | Prudence | Auvry | Sycorax ]

Zam, assuming your bow damage stays within a relevant range and you remove outliers you can probably use a normal probability distribution formula to get within 95% certainty rather easily. I'm no statistician, but it shouldn't be too hard to determine it using that.

Nice work.

Reksar is correct. The best way to calculate your average damage would be to assume a "Normal" (Gaussian) distribution.

First, look only the shots that hit (eliminating head shot and DD proc damage, if you like). Compute a mean and standard deviation for the damage from all of your hits. The Excel/Open Office/iWork function works, or you can program it into your parsing script.

Compute your hit rate - what percentage of all your shots hit - Exclude the DD procs, but I think you include head shots as hits.

Compute your shot rate - how many shots did you take / how many seconds (total) did your test run.

Convert the mean/std. dev. to a rate -> mean * hit rate * shots/second = mean DPS .... std.dev. * hit rate * shots/sec = Error
Your DPS number would then look like: mean DPS +/- Error

_________________
Nikolai Lianfyrr,
Khati Sha Feral Lord.
GM Tailor (250), GM Smith (250), GM Brewer (250), GM Potter (250), GM Fisherman (200), GM Jeweler (250), Fletcher (220), Chef (200)

Interesting results, Zam.

I actually did a similar parse (although at 100% haste) in January of this year and came up with the exact opposite of the results you list.

I have a strong feeling that Raven is correct in assuming the haste rounding on all of these bows significantly changes the picture. I'm (almost) intrigued enough to run another full parsing at 125%!

_________________

Thanks Niko, that was exactly what I was looking for.

Ader, can you post your results, or are they for Tem eyes only? And when you say 100% haste, I assume you mean 41% worn haste + 68% haste of Speed of Vallon; wouldn't that be 109% haste?

Probably he means 41% worn haste plus warsong of zek (60%) haste, but either way the non-v3 haste cap is 100%, so it doesn't matter much; the only two meaningful snapshots of melee DPS are going to be at 100% (however you reach it) and 125%, no?

Because I don't really know how the fleeting quiver works (does anyone?), it's hard to make guesses about how parses at lower haste values will scale up to normal raid haste. I also wonder what the impact of eagle eye is, and whether it doesn't actually affect some bows more than others.

_________________
Ryoku Wing, Swift Tail of Temerity

Soft-skinned alt scum:
[ Ravenwing | Syovan | Prudence | Auvry | Sycorax ]

These were all done on a lvl 55 mob (no headshots) with no buffs besides Warsong of Zek to cap haste at 100%, and with worn ATK capped at 250.

The parses were all pared down from their original length to 12 hours and accuracy hovered around 61% - 62% for all four bows.

Obsidianwood - 166.26 dps,
Primordialwood - 156.79 dps,
Stonewood - 153.55 dps,
Featherwood - 149.98 dps,

_________________

Well now im just confused

_________________
Duke Hendo <65 Pally>
~ Member of Destiny ~

Hendo's Soldiers
Crak Deeler <Enchanter>
Dragonborn <Druid>
Rukus <Wiz>
Ruffio <Wiz>
Zeuss <Shaman>
Matrex <Cleric>

It should be noted that I did not remove proc DPS from the parses. Partly because I'm lazy, but mostly because I don't know enough about shell commands or anything similar that would make removing the relevant text easy. So there's probs some small inflation. Also, they were all rounded down to 10 hour parses, since some of them were 11 hours long.

You figure 2 procs/min at 10 hours is 1200 procs total. There's also the crit chance (7% I think?), making maybe 84 of those procs worth double. 3.92 inflated DPS for the Obsidianwood and 2.67 inflated DPS for Featherwood.

Relatively insignificant, but it's obviously another moving cog in the machine, if you get my meaning.

_________________
Delorne / Oogliest / Kelorekdar / Draesion / Darganor / Morigill / Akkirus

Did all participants use fleeting quiver over the course of testing?

_________________
Faults - 65 ranger of <Temerity>
Beardog - Melee Shaman box
Bahnanas - Enchanter Box
Whirleebird - Cleric Box
Hijinx - Bard box

It was the same ranger testing 4 different bows. Quiver is a given.

_________________

Proud member of Temerity
GM Fletcher (250) Jewel Craft (234) Brewing (226) Smithing (224 + 15%) Pottery (234) Baking (246) Pure Hell (226)
Wulfgar - Lvl 52 Barbarian Warrior, LFG in PoFire!
chown -R us ./base

Interesting. I now understand that my parses were a waste of time since I didn't have 100% haste, which seems unfortunate. But you live and learn, I guess!

Thanks Ader for sharing your data. Once I get max attack on my ranger I may try to duplicate your results with Warsong (and in the process calculate a standard deviation like Niko suggested). Which level 55 mob did you use? I wasn't aware there was a better parsing mob available than the Katta Banker. Also, it is a shame you didn't include a figure for the Agnarr or TZ bows, since they may (?) perform in a superior fashion to 4 elemental ones.

Lack of + to archery on tz/agnarr might hurt.

I was curious if zamiel had used fleeting quiver in his test - since there are 2 different people testing and 2 separate result groups I was curious about all variables.

_________________
Faults - 65 ranger of <Temerity>
Beardog - Melee Shaman box
Bahnanas - Enchanter Box
Whirleebird - Cleric Box
Hijinx - Bard box

Perhaps you should read the OP again. ;)