[19.7 Ground Combat] Smoke Bugs

[kabill]

A couple of smoke related bugs:

1) On the turn you throw a smoke grenade, the sight penalty does not apply immediately. This applies to both player controlled units and aliens. Given how the fog of war works, it's actually most noticeable with the aliens, who will still take reaction shots on the turn it is deployed. I *think* that if a unit is not already in LOS of an alien when smoke is deployed, then the smoke will block LOS properly if a soldier then moves to where LOS would have been (but I've not tested this).

2) I'm fairly certain that the to-hit penalty for smoke isn't working. I've been taking a number of shots today through smoke where the % chance to hit is minimal (intending to suppress) and getting hits anyway. In one instance, I was given a 0% chance to hit with burst fire, and two out of the three shots hit:

I might try and do a proper test of this, but it seems like the penalty is being applied to on the UI but not in the actual attack roll.

3) I've posted about this before, but thought it was something else than it is. When a soldier is kneeling, the UI will not allow a soldier to take a shot through smoke. Standing them up allows them to shoot as normal.

[kabill]

Did some testing to see whether the smoke penalty was working. The short answer is that it does appear to work, but only sort of.

Tested using two soldiers. One was positioned straight on from the target, the other was positioned slightly to the side. I tested first without smoke to produce a base line, and then with smoke.

Test 1: No smoke:

Solder 1 has 22% chance to hit. Out of 100 shots, misses 68 and hits 32 (32% hit)

Soldier 2 has 18% chance to hit. Out of 54 shots, misses 45 and hits 9 (20% hit).

Sadly don't have stuff available to test whether these are statistically significant differences (soldier 2 definitely won't be, mind).

Test 2: With smoke:

Soldier 1 has a 3% chance to hit. Out of 102 shots, misses 87 and hits 15 (15%)

Soldier 2 has a 1% chance to hit. Out of 65 shots, misses 59 and hits 6 (10%)

Obviously, the actual % hit are much greater than that expected. But they're still lower than non-smoke accuracy (based both on hit chance and the actual number of hits scored).

I'm aware there's problems with the method here and without more test shots/different scenarios these tests are not conclusive. But this combined with anecdotal experience I think offer a reasonable indication of something being wrong. Something I do note is that, if smoke is applied as a -30% modifier (i.e. accuracy*0.7), the number of hits scored through smoke is close to the non-smoke accuracy values *0.7 (22*0.7=15.4; 18*0.7=12.6). Perhaps the UI is applying a 0.7 modifier for every smoke-filled tile the shot would pass through, while the hit calculations only count it once?

---

On the issue of kneeling and smoke, here's a screenshot:

Soldier 1 from the previous tests is now kneeling (this is the only change). She is no longer able to shoot the alien. I had a screenshot of the other soldier too, but it won't let me upload. But it was the same.

Interestingly, it was possible to get *some* LOS through the smoke (again, can't upload screenshots). It was being capped at about 4 squares through the smoke. If I moved a soldier into the middle of the smoke and knelt, it would let me take a shot at the alien. So it looks like kneeling + smoke caps the number of tiles you can shoot (and see?) in some way.

[kabill]

One last test on the aim issue. Modded the AR to 1TU burst shots to increase the sample size/time spent testing.

Test 1: Without smoke

Had this time a target with no cover. Displayed hit chance was 37%. Half the sample were taken at an angle, half were taken dead-on (there was no meaningful difference between these). Out of 394 shots, 148 were hits (37.6%) and 246 were misses.

Test 2: With smoke

Shooting the same target from same positions. Displayed hit chance was 6%. Out of 396 shots, 143 were hits (36.1%) and 253 were misses.

I'd say a sample size is large enough to have some statistical power now, and I'm sure that if you calculated it there would be no significant difference between the two percentages. Smoke definitely isn't having anywhere near the effect it should be having, and quite possibly is having no effect at all on hit %.

EDIT: Or it's a problem with the RNG, of course!

EDIT 2: Or not with the RNG. Just tried a similar test, with 6% shots against an alien in cover rather than behind smoke. 200 shots, 20 hits and 180 misses. Still slightly more than expected, but a far closer than with the same hit % through smoke. As an aside, I do note that the actual number of hits does tend to exceed the % chance to hit the UI displays, though.

Edited August 18, 2013 by kabill

[Ol' Stinky]

Thanks for taking the time to do those tests! You're a hero in my book.

[Matthew]

Nice job!