Pistoleer Archive
Thread: Weapons Accuracy Test
For the heck of it, a guildie and I ran a fairly extensive series of tests to measure the effects of accuracy. This first post will contain the data; follow-up will include some analysis.
Range: 60m. We never moved through out the experiment, using just posture changes and /tumblewhilestanding. We both did several posture changes and then waited a few minutes after getting into position to make sure the position was sync'd properly. The distance in the targetting screen never changed.
Assumption: the range mods for weapons is piecewise linear. The quality of the weapon does not affect accuracy.
Weapons used:
CDEF rifle, +0/+42/-80 at 0/35/64m (calculated modifier at 60m: -63.2)
CDEF pistol, +0/+40/-80 at 0/15/64m (calculated modifier at 60m: -70.2)
I was the shooter. I am a master marksman/rifleman/squad leader. We were grouped for the entire test except for the last trial. My pertinent modifiers are: +112 rifle accuracy, +60 pistol accuracy, +30 group ranged defense. My target's pertinent modifier is: +6 ranged defense. She had no dodge, evade, counterattack, etc. modifiers.
There were 8 trials. They are listed below in this format:
(weapon,shooter status,target status)
(total shots, hits, misses -dodges, evades, counterattacks)
1: rifle, standing, standing
285,191,80 - 6,0,8
2: rifle, standing, prone
297,176,120 - 0,0,1
3. rifle, standing, standing+tumbling
222,140,81 - 1,0,0
4. pistol, standing, standing
189,83,102 - 3,0,1
5. pistol, crouching, standing
230,128,98 - 1,0,3
6. pistol, standing, standing + tumbling
182,69,112 - 1,0,0
7: pistol, standing+bodyshot2, standing
96,62,29 - 0,2,3
8: pistol, standing, standing (ungrouped)
205,116,79 - 5,0,5
Some interesting tidbits:
* Ranged Defense -works-. Or at the very least, the group ranged defense mod SLs have works. The hit rate went from 56.6% to 43.9% and the miss rate went from 38.5% to 54.0% (the rest are D/E/Cs). This is roughly a 15% swing for +30 ranged defense. The trouble seems to be that it's very difficult for people to collect large amounts of +Ranged Defense since it's given out in rather small chunks.
* Active defenses -seem- to be tied to the hit rate. In particular, more hits, more active defenses seem to occur. This would imply that active defenses act as you would think, in that they prevent a hit rather than making misses more likely (sort of a subtle distinction). In particular, roughly 3% of potential hits seem to be defended by a +0 defense.
* The previous point implies that it makes more sense to talk about -miss rate- than hit rates. The theory is that the game checks first for a hit/miss, then each of the actives in turn.
* Tumbling seems to prevent counter-attacks. Or it might just reduce the chance of all active defenses. There were 0 counterattacks during the tumbling portion, and only 2 dodges (about 0.5% occurance) compared to approximately 3% counterattacks and dodges everywhere else. But the numbers involved are small enough that it's hard to tell.
* Evade only occurred when being attacked by a special move. There were no dodges when being hit by a special; this might imply that evade is just a dodge that occurs with a special move. The message you get from a special is different from a regular attack: when a regular attack is actively avoided, the damage is -not- listed. When a special is actively avoided, the damage -is- listed. So it could just be inconsistent use of the word. Or it might just mean we were unlucky on dodges that series (would have expected 3 out of 96, however).
* Being prone -may- affect the active defense rate (not enough data here, though - but active defenses dropped from about 5% of all attacks from standing to 0.3% -one counterattack- when prone)
* Special moves most certainly affect hit rate. Bodyshot2 let me hit the target 64.6% of the time (30.2% miss rate) while regular attacks with all other variables held constant hit 43.9% of the time (54.0% miss rate). That's roughly a 14% swing, which seems to work out to something like +28 accuracy (a very rough estimate seems to be +2 accuracy = +1% chance to hit - it's probably not linear in that manner, though). In this particular case, using BS2 over a regular shot improves the DPS by almost 50% just from accuracy - never mind the damage bonus. This potentially has huge ramifications if some specials have a much larger accuracy bonus. It may mean the damage multiplier and the delay multiplier isn't the end-all, after all (maybe double-tap already IS the long-ranged, accurate special...).
* There's surprisingly little effect from +accuracy. I have +112 rifle, +60 pistol, AND my pistol had a worse range modifier, and my miss rate went from 28% to 54%. That's about a 50% increase in hits for my rifle compared to pistol; actually, I suppose that's not so little. It's like getting BS2 accuracy added to all my rifle shots. Still, it's not enough to be immediately apparent in casual use.
* Tumbling added about 8% to the miss rate. Target being prone added about 12% to the miss rate. Shooter being crouchedsubtracted about 11.5% to the miss rate. This might imply that being crouched/prone helps your accuracy more than it helps your defense. Or it might just be because the base miss rate was lower during the prone test than the crouched test. Upon reflection, I wish I'd run identical tests for pistol and rifle.
* I will always be grouped while fighting from now on.
If anyone else has controlled test data, please post so we can figure out how the mods really work. I'll put up some conjectures based on my datatomorrow after I get some sleep.
I would be interested in seeing a test serieswith you standing/kneeling/prone with rifle and then the same with pistol/carbine vs. a standing target.
The reason being I think we need to determine if +accuracy mods cap at some point - ie what is the "best" hit/miss ratio achievable? One of the things that leaps out from those tests is that in the base stand/stand tests, a +52 accuracy and -7 weapon mod differance only combined for a 15-16% differance in actual hit percentages.
Some very interesting tidbits in there. Thank you
Excellent study. I love this kind of stuff.
It would be interesting to see some numbers atthe inaccurate extreme. For example a raw noob with novice marksman using the same rifle at 60m against a prone target with a decent ranged defense (perhaps grouped with a SL too). In factI wonder if there ever comes a point where you basically cannot hit...
More data (Set B). The target is a different player (only +2 ranged defense), but the range is the same (60m) and I used the same rifle I used in the first test.
Assumptions: The target accuracy mod (the one that shows up left of the HAM that shows above the target) means something. Aim modifies the target accuracy mod by a constant amount. I used aim to circumvent the problem of the accuracy mod capping at -50 to establish that (assume +0 is shooter+target standing):
Shooter crouch: +15
Shooter prone: +30
Target crouch: -5
Target prone: -25
1: Stand, Stand + Grouped
318,213,95 - 5,0,5
2: Prone, Stand + Grouped
185,176,8 - 0,0,1
3: Stand, Stand (No Group)
276,214,39 - 8,0,15
4: Stand, Prone (No Group)
325,230,87 - 5,0,3
Errata: I forgot, but in trial 6 in the first experiment, the target's tumble wore off somewhere near the end of the trial. So the miss rate for trial 6 is a bit higher than actually listed. Also, when doing test 4, the target had to stand up to heal multiple times (I stopped firing when this occurred). However, this kept changing the range slightly (about 1m every 2-3 stand/prones). I had to make microadjustments. Early on, I made a few microadjustments -while firing- which obviously affects accuracy. When I realized this, I used the breaks to make the adjustments. This should only be a few shots, but it does mean the miss rate in B4 is slightly lower than indicated.
(A6 - miss rate is higher than listed, B4 - miss rate is lower than listed)
Analysis follows...
Great set of posts Noules000!
Im looking forward to the rest of your tests
Hawks - Flurry
Great thread.
Some people have been saying suppression took an accuracy hit along with the nerf on it, could you possibly test it and head/leg shotsas well to see if they measure up to bodyshot?
As a conclusion on just a small portion of your work, would I be correct in thinking that a Master Pistoleer with Ranged Weapon Support 4 from Marksman who has a +13 Ranged Defense (I believe) only gets missed an additional 7.5 times out of 100?
Yay, accuracy tests! Great post! I think we really needed this to firm up our understanding of combat.
BlackPearl wrote:
As a conclusion on just a small portion of your work, would I be correct in thinking that a Master Pistoleer with Ranged Weapon Support 4 from Marksman who has a +13 Ranged Defense (I believe) only gets missed an additional 7.5 times out of 100?
Assuming my understanding is correct - it would be 6.5 times out of 100 (13/2). I don't have a target with lots of ranged defense, so this comes from my experimentation with group ranged defense (I assumed group ranged defense is equivalent to ranged defense). Depending on what the base hit rate is, this could be significant or very insignificant. Something to note though is that ranged defense presumably all stacks. It's possible to get a lot of ranged defense (and melee defense) by taking a line or two of a few melee classes. TKA precision line is the most valuable (+30 ranged defense!) followed by fencer stances&grips. It should be noted that defense is harder to get than accuracy, presumably precisely because it is stackable. The maximum that I could get when working out a defensive build was +120 (90 from self, 30 from a master SL friend).