Again on reaction fire

[DarkRain]

Hmmm.....

1) Telok said it took two more hits to down the bloodspider, which, had already taken two hits, that means that when it was stunned stun damage was at least about 1/3 of its total HP -- assuming the last of four hits made minimal damage the first two should have done nearly 2/3 damage to its HP, so not likely.

2) Yes, but robotic actors have hardcoded immunity against this (stun damage, they can still receive normal HP damage)

[Telok]

It was a combat bloodspider, one of the tough ones.

[DarkRain]

@Telok I didn't mean it is unlikely for the boodspider to have taken 4 hits, I had assumed it was a combat model, I meant that if it still had enough HP to need two hits to take it down, it is very unlikely that it received enough stun damage to stun it, considering that with 999 resistance to stun it would receive only 1 stun per hit, that is unless there is a bug, which is what I'm thinking (or maybe you modded in a stun weapon with 1000+ damage that you aren't telling us about )

[Anarch Cassius]

There was a commit change recently addressing a RF bug that may explain the aliens seems to move and fire for full and then take full RF later. Hadn't seen an announcement and I know this irritates the crap out of people so I thought I'd mention it.

[Telok]

I've updated to the Dec 18 nightly and found something I absolutely did not expect.

I ran another pair of sets at square 45 with assault rifle RF using single shot. They took 7/494 RF chances at 11% to 15%.

Then I tried burst fire RF.
At square 45, 8% to 11% hit chance, 28/30 RFs
At square 55, 5% to 7% hit chances, 202/341 RFs
At square 65, 4% to 5% hit chances, 22/48 RFs

So, um, weird. At a lower %hit and costing more TUs the soldiers RF 20 times as much. I guess this explains the MG numbers to a point. The more shots that will be fired means the more likely an actor will RF once the TU requirement is met. And soldiers using ARs won't RF if an alien is blocking LoF, even though soldiers with a sniper rifle will.

I'm not modding anything and definitely not nuclear stun ammo.

[Telok]

I played a bit with MGs and SMGs.

Your chance of getting off an RF shot is proportional to the number of shots you will fire. Soldiers with MGs will fire on every single RF provocation at 70+ squares of distance using full auto while standing. They will not hit anything. The same thing happens with SMGs except that you have to be in targeting range. Based on the MGs (I have to go back and confirm it) it looks like you need a minimum 1% reported chance to hit to take a RF.

So this is what I know so far.

RF is based on:
1) Being in targeting range, but closer is better
2) Clean line of fire (green targeting line) except for through_wall weapons
3) Enough TUs spent by the target to provoke RF
4) The number of shots the RF will fire
5) Maybe a >=1% hit chance

I may need to start modding weapons if I want to find exact numbers but the trend is obvious, more shots = more RFs.

[Telok]

I've played with the plasma rifle this time, because it has three RF modes.
Everything in this post is at a range of square 68.

Standing Auto-fire, 1% reported hit rate, six tests run.
1) 16 out of 150 RFs taken.
2) 4 out of 102 RFs. 10.6%
3) 2 out of 97 RFs. 3.9%
4) 5 out of 67 RFs. 7.4%
5) 1 out of 187 RFs. 0.5%
6) 24 out of 148 RFs. 16.2%

Notes: On tests 1, 2, and 6 the hovernet bugged and moved around invisibly while it's model remained in one place. On tests 1 and 2 this happened on the 11th turn of combat (13th turn since starting the map) and the soldiers stopped taking RFs at it. On test 6 it happened on the third turn of combat but the soldiers continued to RF until the 10th turn, no more RFs after turn 10. Also interesting is that test 5 had only one hovernet to shoot at while test 6 had two hovernets and four other aliens to soak up ammo. Soldiers using auto will try to RF through blocking aliens.

Crouched snap shots, 4% to 6% reported hit rates, three tests.
Zero out of 496 RF shots taken.
Depressing.

Crouched burst fire, 2% to 4% reported hit rates, three tests.
1) 3 out of 206 RFs. 1.4%
2) 0 out of 101 RFs. 0%
3) 4 out of 106 RFs. 3.7%

Notes: Test 1 had it's first RF in the second round, the last RF in the last round (24th of combat), and the middle RF about in the middle of he combat. Test 3 had all four of it's RFs in the very first round. Interestingly test 2 had only one hovernet to shoot at while test 3 had two hovernets and two other aliens.

Next I abandoned Phalanx soldiers and selected a team of Tamans (off topic: a squad of hovernets with miniguns equipped looks hilarious), these buggers have 75+ accuracy and 80 to 90+ assault skill. This makes a difference.

Standing auto-fire, 2% to 4% reported hit rate, two tests.
42 out of 47 RFs taken.
An 89% RF rate, utterly amazing. Even better is that they actually hit once in a while.
Plus they tried to shoot through blocking enemies again.

Crouching snap shots, 11% to 13% reported hit rate, 4 tests.
1 out of 682 RFs.
It's infinitely better than 0, but that's not saying much.

Crouching burst fire, 6% to 8% reported hit rate, 6 tests.
1) 27 out of 47 RFs. 57.4%
2) 19 out of 19 RFs. 100%
3) 6 out of 114 RFs. 5.2%
4) 21 out of 44 RFs. 47.7%
5) 11 out of 17 RFs. 64.7%
6) 11 out of 51 RFs. 21.5%

Notes: Tests 2 and 3 are when I really started to realize that the target environment may be having an effect on RFs. I was double checking my scratch pad when I realized that test 2 had two hovernets, a shevarr with an assault rifle, and three other aliens. Test 3 had a single hovernet. Tests 4 and 5 had s single hovernet and three filler aliens. Test 6 had a hovernet and two filler aliens. Again they shot through blocking enemies.

Conclusions: Actors with 30s skills RF far far less than actors with 80+ skills, no surprise there. The number of shots fired is vastly more important than the accuracy of the weapon. Multi-shot RF modes will try to shoot through a blocking alien, I need to try with the particle rifle or something to confirm single/multi through alien shooting. And I need to do follow up testing on the target environment because it looks like a target rich environment gets more RFs than a single target.

[tembero]

Any way of added a 1 tile movement mechanic to the reaction fire, I.e. After you take your reaction shot your soldier moves to a previous designated square. I notice that on the aliens turn a marker appears if you try to move your soldiers showing the path they would take perhaps when selecting reaction fire giving the option to click a point on the map to retreat to would be like leaning around the corner. at least giving the reaction shot a chance to seem like you have "stunned the alien for 2 T.U's?

[Denthar]

There was a commit change recently addressing a RF bug that may explain the aliens seems to move and fire for full and then take full RF later. Hadn't seen an announcement and I know this irritates the crap out of people so I thought I'd mention it.

If it was committed then it's not fixed as of the Dev 2.5 build for Jan 10th, as I'm experiencing quite a high frequency of aliens happily moving sizable distances, firing off shots, and then hammering my squad with RF come my turn.

It is extremely irritating, as you've noted, as it's incredibly unfair and unfun.

Without diving into the code, I wonder if aliens have their TU replenished during the human turn (therefore allowing them have a massive amount of RF) rather than during their own turn.


As an aside RF seems to have other issues in this game, it might be a design choice, but I find it extremely odd that aliens don't RF often if you move about in their LOS, but go RF crazy when you try to shoot (often killing your squad members before they can complete their action).



(Nb, first post. Be gentle )

[Patupi]

It's based on the TU cost for whatever you're doing. Since moving is only 2 to 3 points (more for awkward terrain) and firing is 4TU and up then moving often won't trigger an RF response. From what I understand it's based on view. If an alien has you in view and has a weapon set for snap RF that takes 8TU to fire, you need to spend 8+1, 9 TU in view at one go. Dodging out of sight and back into sight resets this. But fire, often goes over this limit and the alien gets RF fire. If the weapon mode their set to requires more TU to fire you can spend more time in view before needing to dodge out of view again.

[H-Hour]

FYI, as of yesterday, the RF threshold has been removed in 2.5-dev. Reaction fire should no longer be prevented from firing if the chance to hit is too low.

[Xeinar]

H-Hour, did the team take into consideration the idea of not resetting the count when a unit moves between covers? This would trigger RF more easily, and would make more sense (if I see someone moving into cover, I know he'll come out from there).
I think that if the count is not resetted, the RF would seem more natural.

[Duke]

We have considered it, but it's technically (close to) impossible.
For the code, it's visibility on/off. So the code can not distinguish between
a) walking behind a broken fence and
b) dodging behind a crate and
c) walking into a house and out of it through the backdoor.
The latter is certainly a non-RF situation imho.

[H-Hour]

I'm not entirely convinced by that rationale. I struggle to think of a single map we have where (c) could actually happen within a single soldier's field of view. (a) and (b), however, can happen very frequently on some maps.

[Xeinar]

You know the code, I don't, so most probably you are right  , but what about creating a number of counters for each unit in game? I mean, each of my soldiers will have as many counters as the aliens in play (and viceversa). While in visibility, the counter add the TU spent (triggering, in case, RF). At the end of the turn the counter is resetted.

From a logic point of view is it not so hard to implement but again, I'm a profane with no knowledge of your work so far