I decided to do some parsing with 3 weapons of nearly equal ratios to test the effect of weapon delay and ratio on DPS. I used one weapon at a time for accurate results, opting to use H2H for the opposing weapon slot for easy identification of damage of the tested weapon while maintaining a realistic environment by not using a non-melee item. The parses are all cut down to exactly 7 hours (25,200 seconds) as comparison parses on a level 98 Test dummy with 20 upgrades to mitigation. The three weapons tested wer The Dagger Pernicious, Midnight Star, and Eclectic Dreamblade. All augs remained constant in respect to weapon slot, with mainhand piercers having Blessing of Might and offhands having Slime Coated Pebble of the Deep, which is incorporated in the listed weapon stats below. I had full raid buffs with no poisons or other proc buffs or short duration buffs except Thief's Eyes. Listed directly below are some observations from the parses, followed by data and further explanations.

1) Weapon delay relates directly to the frequency of Massive Strikes.
2) A piercer with a higher damage to delay ratio but lower delay than another piercer can actually do significantly less DPS due to less Massive Strikes.
3) A piercer with less delay than another piercer with the same BS DMG has significantly less BS attempts and BS DPS due to less Massive Strikes

Weapons tested in Mainhand:
The Dagger Pernicious (103 Dam / 21 Del = 4.90 Ratio, 51 BS DMG, 320 DD Proc)
Midnight Star (94 Dam / 19 Del = 4.95 Ratio, 51 BS DMG, 0 DD Proc)

Weapons tested in Offhand:
The Dagger Pernicious (101 Dam + 1 Cold Dmg / 21 Del = 4.82 Ratio, 320 DD Proc)
Midnight Star (92 Dam + 1 Cold Dmg / 19 Del = 4.86 Ratio, 0 DD Proc)
Eclectic Dreamblade (115 Dam + 1 Cold Dmg / 24 Del = 4.80 Ratio, 192 DD Lifetap Proc)

Mainhand
The Dagger Pernicious: 3858 DPS (2012 Pierce, 1797 BS, 49 DirDmg) - 64664 Pierce attempts, 13655 BS attempts, 3685 Flurries, 968 Massive Strikes
Midnight Star: 3546 DPS (2019 Pierce, 1527 BS, 0 DirDmg) - 71225 Pierce attempts, 11453 BS attempts, 3958 Flurries, 887 Massive Strikes

Offhand
The Dagger Pernicious: 1540 DPS (1516 Pierce, 24 DirDmg) - 49956 Pierce attempts, 1234 procs
Midnight Star: 1535 DPS (1535 Pierce, 0 DirDmg) - 55206 Pierce attempts
Eclectic Dreamblade: 1502 DPS (1486 Slash, 16 DirDmg) - 43716 Slash attempts, 1292 Procs

I also parsed a Rusty Mining Pick with a delay of 38 to test the potential theory that Massive Strike is firing off of weapon delay instead of merely backstabs. Here are the results:

Rusty Mining Pick: 735 DPS (167 Pierce, 568 BS, 0 DirDmg) - 35488 Pierce attempts, 11255 BS attempts, 1924 Flurries, 899 Massive Strikes

The Rusty Mining Pick with the largest delay had the least number of attempted backstabs, but more Massive Strikes than the Midnight Star. It still didn't come close to matching the Dagger Pernicious in number of Massive Strikes.

The Midnight Star has more piercing damage than the Dagger Pernicious, but the lack of a DirDmg proc puts it below Pernicious in overall DPS. However, it would seem that the Midnight Star could jump ahead again when modified by various buffs, such as shaman epic, war cry, auspice, etc., as the piercing damage would increase as the proc damage stays the same.

Here is a list of the number of single, double, triple, etc. backstabs for each piercer, along with the total % of the total attempted BS's, in an attempt to find why the Pernicious had so many more attempted backstabs than the other piercers:

The Dagger Pernicious, 4660 BS's (1= Single BS, 2 = Double BS, etc.)
1: 29 - 0.62%
2: 1113 - 23.88%
3: 2851 - 61.18%
4: 519 - 11.14%
5: 122 - 2.62%
6: 22 - 0.47%
7: 3 - 0.0006%
8: 1 - 0.0002%

Midnight Star, 3884 BS's (1= Single BS, 2 = Double BS, etc.)
1: 13 - 0.33%
2: 933 - 24.02%
3: 2331 - 60.02%
4: 484 - 12.46%
5: 99 - 2.55%
6: 18 - 0.46%
7: 6 - 0.15%
8: 0 - 0%

Rusty Mining Pick, 3818 BS's (1= Single BS, 2 = Double BS, etc.)
1: 16 - 0.42%
2: 947 - 24.80%
3: 2241 - 58.70%
4: 482 - 12.62%
5: 106 - 2.78%
6: 20 - 0.52%
7: 4 - 0.10%
8: 2 - 0.0005%

This data seems to be very consistent. I just can't seem to figure out why the Dagger Pernicious has so many more total backstabs and Massive Strikes than the others.

Any thoughts? If nothing else, it was still an interesting parse session.

-----------------------------------------------------------------------------------------------------------------------------

I feel confident with the above offhand calculations. However, below are the results of some 12 hour parses of the Midnight Star and Dagger Pernicious mainhanded, using Frenzied Stabbing for more backstab attempts.

Mainhand
The Dagger Pernicious: 6170 DPS (1994 Pierce, 4128 BS, 48 DirDmg) - 110316 Pierce attempts, 53098 BS attempts, 6213 Flurries, 4052 Massive Strikes
Midnight Star: 6207 DPS (2035 Pierce, 4172 BS, 0 DirDmg) - 121852 Pierce attempts, 53854 BS attempts, 6823 Flurries, 4246 Massive Strikes

The Dagger Pernicious, 18038 BS's (1= Single BS, 2= Double BS, etc.)
1: 91 - 0.50%
2: 4323 - 23.97%
3: 10824 - 60.00%
4: 2258 - 12.52%
5: 423 - 2.35%
6: 95 - 0.53%
7: 20 - 0.11%
8: 4 - 0.0002%

Midnight Star, 18222 BS's (1= Single BS, 2 = Double BS, etc.)
1: 76 - 0.42%
2: 4409 - 24.20%
3: 10832 - 59.44%
4: 2251 - 12.35%
5: 514 - 2.82%
6: 106 - 0.58%
7: 28 - 0.15%
8: 6 - 0.0003%
9: 0 - 0%
10: 1 - 0.00005%

A very interesting find, thanks for sharing it.

So in theory proc rates are normalised by weapon delays (but I suspect things like HH effects aren't considered and bias this somewhat).

The Massive Strike numbers have a large disparity - 9% higher. Your delay change from 19 to 21% is approx 10% slower - similar at least to the MS change. However, massive strike is procced off backstab and backstab hasn't changed in delay. Maybe we're seeing MS rate being modified by weapon delay to try and keep a constant proc rate, but not taking into account the rate comes from BS delay and not weapon delay. That's likely a bug if so.

I'd lilke to see some slower piercers tested.

http://eqitems.13th-floor.org/itemse..._4=&next _4=0

maybe for ideas. :-)
Edit: Looks like Rusty Spear is a good candidates.

Shouldn't it have been the other way around as the MS is FASTER than the DP??

No, the theory for the procs is : the slower the delay is the more you have chances to proc to have an uniformed proc rate between weapons. So in this case BS doesnt change but a slower weapon "procs" more massive strike on a succesful BS than a quicker one, same rules as the DD procs. It's likely a bug considering BS speed isnt modified by weapon delay making slower weapon better for massive strikes than quicker one.

I'm parsing a Rusty Mining Pick now, and will have the results up later this afternoon.

It has been proven through testing in the past that the higher delay a weapon has, the more often it procs, which can also been seen from my data above comparing two weapons with the same procrate, the Dagger Pernicious and Eclectic Dreamblade.

Confusion over MS = Massive Strike and MS = Midnight Star

I did a couple very hasty parses with some more extreme weapons, just to see what the impact was. I used Frenzied Stabbing to increase the backstab rate dramatically, just to aid parsing. Maybe this has an effect though on MS rates. (Needs verifying)

Locustlure: 17 delay, 1 hour
5169 BS attempts (including misses and MS hits)
456 Massive Strike

Rusty Shortened Spear: 32 delay, 1 hour
5127 BS attempts (including misses and MS hits)
404 Massive Strikes

Both with chanter haste + overhaste clicky. If it's genuingly using proc mechanics removing haste may have an impact on the proc rate too.

Edit: I forgot to remove worn haste too, but with 40% haste instead of 115% haste:
Rusty Shortened Spear: 32 delay, 1 hour
3221 BS attempts (including misses and MS hits)
275 Massive Strikes

So there are differences, but these are relatively small sample sizes. It's also in the opposite direction to Sindier's longer parses above. Fewer MS with a longer delay weapon here. I think I need longer parses, but it's not obvious even given excessively different delays. Also regarding your parses Sindier:

So why more backstab attempts with pernicious vs star? It's clear there are ~90 more massive strikes, but we see 2200 more backstab attempts. The delay shouldn't impact on that, so what did? You say the time was cut down to the same duration, so it's not that. Do the parses indicate anything else? Eg higher rate of triple attack? Some blind luck somewhere? I assume attempts include misses so accuracy fluctuations shouldn't break it either. Any bad lag experienced during the parse? Backstabs rounds can be entirely eaten by lag.

I edited my original post with more data. The number of triple BS's, etc., seems to be consistent with the tested daggers, as well as the added parse of the Rusty Mining Pick with large delay.

I'm still stumped as to why the Dagger Pernicious has so many more total backstabs.

It's a conundrum. Do you have time/willing to retest Penicious, just incase of some accident?

I can't explain it. Maybe there's something specific at 21 delay too. Odd though.

omg lol

Anything constructive to add?

Sindier, I did some stats and based on 10566 backstabs with a 9.226% chance of proccing massive strike (so 887 MS giving combined 11453 backstabs) there's a standard deviation of +/- 31.

Ie 5% of the time measuring such a quantity we'll see figures < 825 or > 949. There'll be a similar s.d. for the 968 flurries from pernitious too giving a substantially overlapping distribution. Realistically it's unlikely that this is down to chance, but it's not sufficiently unlikely to claim the difference is significant.

In other words - we need longer parses. I'd suggest using frenzied stabbing to increase the rate of gathering data.

I wouldn't mind seeing how a player made, non-proc weapon would parse in these scenarios. Take Extravagant Infused Serrated Dagger (34/21) and see what you get for counts, if you want

Thanks for the input, informative as usual!

Very strange, it would seem if these #'s are right that Dagger Pern will outdps Mid Star, which me and the 20 or so other poor souls without Leypse simply didnt see possible.

Perhaps its because of the difference in inherent procs, ones a DD and ones a de-aggro, not that that makes any sense... or maybe its like you said and has something to do with friggin 21 delay... these #'s are weird though. Very interested in seeing this parsed more.

Years ago, wasn't there a bug rumor that 21 delay was actually the same as 19?

If so, then maybe old coding affected the 21 delay on Pern, and it delays only 19.

That rumour had a base in reality, but it was never that extreme. Specifically the delay of weapons was rounded into discrete quanta, 1/12th of a second IIRC - whatever an Uzun turned out to be anyway. This meant that certain weapons delays didn't speed up if you used a higher overhaste rank, or given the same overhaste rank sometimes two neighbouring delays ended up being the same.

In the end the devs applied a fix where the delay error was accumulated and when it amassed to an entire extra attack round the extra attack was applied... or something akin to that anyway. Ie it got fixed.

This could still be nothing more than luck. It needs more data.

I'm pretty confident in the offhand parsing, as the DPS for both weapons was very constant for the last 1-2 hours of the parse. The annoying thing about testing the backstabs is it ties up my computer for the duration of the parse =) I'm going to do 2 longer parses with the two offhands though, and I'll add the data to the original post.

New numbers are posted up on the initial post, and is closer to what I expected. However, now the Midnight Star has more backstabs! It looks like Massive Strikes do seemingly proc from backstabs and not directly from weapon delay, as the number of MS's per minute increased to over 5 per minute compared to 2 per minute without Frenzied Stabbing. I don't have time to further analyze it at the moment, but the backstab rates seem to still be constant. If nothing else, these parses will at least serve as a baseline if we get an AA that increases Massive Strike rates or similar with the Veil expansion.

Percentage wise though the difference is far closer. It's also a far larger sample size, due to longer parse and using frenzied stabbing.

I think the previous result was just a lucky streak. The stats claimed it was unlikely, but do enough parsing and even those 1 in 10 or 1 in 20 chances of being incorrect will show up sooner or later.

It was also sort of fun to see a "10 shot" backstab from Massive firing from the huge sample...though the odds of that happening are like winning the lottery.

Higher Delay produces a higher proc rate to compensate for not swinging as often. That's why when warriors use to shakerpage PoFire they use to duel a shaman and get slowed first.