## First test of 2roll question

I decided to make a baseline. As I write this, it’s in progress. [finished – see below]

I’ve got a level 70 nightelf priest using a Calming Spore Reed. She’s got a little crit gear with a paper-doll crit of 6.59%. [next test I’ll do ‘naked’ to remove that question]. She’s shooting at Dr Boom, a level 68 mob (non-elite).
By the book…

If I recall correctly, this works as per spells. If this is a spell, I start with a 98% chance of hitting. That shrinks to 91.41% hit, 6.59% crit, 2% miss (or resist).

By the 2-roll system, the 2% miss (resist) stays. But I get 6.59% of 98, or 6.4582% chance of crit, and 91.5418% chance of ‘regular’ hit.

As I write this, I’m in the midst of testing. Dr. Boom is down to 40%. A quick snapshot gives me 1905 hits, 135 crits, and 40 resists. (Yes, it’s resists. More confirmation of spell.) That’s 2080 total attacks.

By the book (BTB), that’s 1.93% miss, 6.49% crit, and 91.59 hit (rounding gives us the extra .01%).

By the 2-roll, we still get 1.93% miss. Our crit is 135/2040= 6.62% making our hit 93.38%.

Since the crit is the key element… the BTB is off our paperdoll by -0.1%. The 2R is off by +0.03%. (note the odd additional point that my hit is better than paperdoll by 0.07% as well. I’m wondering if weapon skill “shoot” is working in here somewhere.)

I will edit this (by adding after this line) with final numbers. But as of this instant, the case for wand being 2-roll is strong.

Final numbers

• Hits 2712
• Crits 189
• Resists 58
• Total attacks 2959
• Total damage attacks 2901
• Resist rate: 58/2959 = 1.9601%
• Crit rate BTB = 6.3873%
• Crit rate 2R = 6.5150%

(remember, paperdoll crit rate 6.59%)

Log notes I started it when the target was at 92% health, so it’s got a slight possible data lag. Still, with 2700 hits I’m getting a decent confidence level, and the tentative conclusion is:

All spells are 2-roll tables (at least).

Next test will be “naked”. I’ll strip all gear, and use a wand that doesn’t add to my int (crit) level. I want to see if the anomalies (1.96 vs 2% miss, both crits below paperdoll) continue — if so, there are even more errors in the ‘presently known’ tables to be resolved.

~ by Kirk on September 5, 2007.

### 7 Responses to “First test of 2roll question”

1. Just left a reply on BRK’s blog but I’ll put it here as well: nice work! I think your conclusion is right, although more testing cannot do any harm :-).

But a question keeps nagging me: does this matter? Is there a situation where knowing whether your attacks are on a one or two roll system will help you decide which gear to choose?

2. Bastiaan – in a word, yes it matters. Since I started to get longwinded, I’ll change it to a post to explain why.

and thanks for the compliment.

3. Btw, may i sugeest using a low level enchanter / AH wand with no +int, and relatively low dps, with a fast speed. That will give more data faster, and Dr B wont die so fast

4. errgh – not die so fast? 3000 attacks at roughly 40 attacks per minute was FAST? grrgle….

Seriously, I know what you mean. My pondering right now, however, is trying to puzzle the effect of LEVEL on crit. Or maybe it’s weapon skill…?

5. How are you sure that wand hits are counted as spells? Just because you get “resist” instead of “missed” doesn’t mean they’re really treated as spells. They still might be on a melee or ranged table, thus not telling us anything about 1 vs 2 rolls for spells.

Unless you’re using level 1 smite or some such and I missed that in the above.

6. Tanilin, good question. The answer is while I know it’s not melee, my confidence it’s not ranged is only high, not certain.

There are two cues that indicate that it’s probably not ranged. First is the cue for a ‘miss’ – resist. The second cue comes from the second test, where my spell crit was just over 3 and my ranged crit was just over 5. In that test, the number of crits – while still somewhat peculiar in my eyes – were significantly closer to 3% than to 5%.

Again I can’t say it definitively, but at this time the clues indicate it to be the most likely situation.

7. […] table. BRK was concerned with how ranged critical hits are determined, while Kirk got into a whole series of posts with an analysis of spellcasting, and a discussion of “why it […]