[OLD] Old OXCE discussion thread

[Yankes]

lasted nights aren't supported. Correct data need by extended is in first post of this thread.

[SophiaThe3rd]

quoted from the first post:

Download available on mod site (require latest nightly):
https://www.openxcom.com/mod/openxcom-extended

perhaps just edit that line?
Great work though! I LOVE all the new fuctionality

[Yankes]

Right, this line is not accurate right now but I plan fix it in near future by making its true again.
Probably most important line in readme is "OpenXcom date:   Nightly 2016-01-02" this is exact version of nightly that Extended was build on.

[HelmetHair]

I have a functionality request.

Auto shot weapons with long bursts increase their accuracy per shot after the first 3 shots by a few percentage points with each shot.

hopefully to simulate correcting or " walking" the fire onto a target.

so as an example you have an LMG with a 8 shot burst starting with a base accuracy of 25% and moving up ...all these numbers are hypothetical. so it looks like

25/25/25/28/31/34/37/40

thoughts?

[ohartenstein23]

@HelmetHair - the increasing accuracy is somewhat doable with the changes to make shotgun pellets configurable in OXCE+.  The old behavior was that the pellets were kind of like an auto shot that decreased in accuracy after each one, and the amount that it decreases can now be set it the ruleset.  Make it a negative number and voilà, increasing accuracy shot, as long as you're okay with no bullet travel time and setting autoShots to 1 with shotgunPellets equal to the number of shots you want.

[ivandogovich]

Right, this line is not accurate right now but I plan fix it in near future by making its true again.
Probably most important line in readme is "OpenXcom date:   Nightly 2016-01-02" this is exact version of nightly that Extended was build on.

Currently, this version of the nightly is unavailable.  Would it be possible to link a copy in the first post?

What I did to get extended running ( and I'm going to produce a video  on this for tutorial purposes unless its not correct ) is:
- use a recent nightly,
- copy in the Data file from the first post,
- copy over original files from steam,
- copy in the latest OXCE executable.

The game launched OXCE when I did that, but I didn't try to check it for bugs (like chryssalid weapon bug).

Please let me know, Yankes, if those previous steps are incorrect.
-

[Yankes]

@HelmetHair - the increasing accuracy is somewhat doable with the changes to make shotgun pellets configurable in OXCE+.  The old behavior was that the pellets were kind of like an auto shot that decreased in accuracy after each one, and the amount that it decreases can now be set it the ruleset.  Make it a negative number and voilà, increasing accuracy shot, as long as you're okay with no bullet travel time and setting autoShots to 1 with shotgunPellets equal to the number of shots you want.

I would personally do not touch "shotgun" mechanic and add new thing to it. It can't have explosive ammunition and flying projectiles. It should be rewritten from the scratch to be usable.

For the feature itself, there is barley difference between flat chance for each shoots and increasing chance for each shoot. Nobody will notice it and over all effective of weapon will be same (of corse if we assume that average chance of each weapons is same).

Currently, this version of the nightly is unavailable.  Would it be possible to link a copy in the first post?

What I did to get extended running ( and I'm going to produce a video  on this for tutorial purposes unless its not correct ) is:
- use a recent nightly,
- copy in the Data file from the first post,
- copy over original files from steam,
- copy in the latest OXCE executable.

The game launched OXCE when I did that, but I didn't try to check it for bugs (like chryssalid weapon bug).

Please let me know, Yankes, if those previous steps are incorrect.
-

You do not need current nightly only data folder from first post.
Simply extract "data.zip" to destination folder where you want have extended version.
Extract exe to that folder. Copy UFO or TFTD to correct subfolders and after that it should work.

[HelmetHair]

For the feature itself, there is barley difference between flat chance for each shoots and increasing chance for each shoot. Nobody will notice it and over all effective of weapon will be same (of corse if we assume that average chance of each weapons is same).

I said the numbers were hypothetical. So what if we increase the chance more, and from every shot from the beginning? that would be noticeable I would think and changes the chanches a little.

25/30/35/40/45/50/65/70

The effectiveness would increase with implementation of autoshot reaction fire and would be noticeable over time if balanced correctly I believe.

[Yankes]

If you have that big step then this weapon become frustrating in use. Not mentioning losing lot of ammo on first shoots.

[HelmetHair]

Are you trying to tell me that you don't like the idea because you think I am asking you to replace the existing auto shot?

That is not the case.

or

is it because of the hypothetical numbers I have listed? they are just a silly example to try to explain the idea.

perhaps I should explain the reasoning better and maybe I should of instead be asking for an additional new fire mode... let's call it long burst.

A light machine gun such as the SAW usually has a very large ammo capacity so wasting ammo is less of a frustration ingame.

With a weapon of this type the idea is a trade off of accuracy for volume of fire. The reason is for a slightly better reflection and viability of common real world tactics like suppressive fire and overwatch.

Suppressive fire, used to deny an enemy vision or mobility and over watch is to protect or warn men from threats they might not see.... I know you know this, but I'm writing it out for my own benefit.

Most of this will be accomplished already through improvements I belive are planned on existing auto fire.

However, why I am asking for an incremental increase in accuracy is that when you shoot a machinegun; the results of your fire can give clues on where to aim to hit what you want.

[Yankes]

I simply think this is not worth effort. Answer one question, what is difference in long run between 10/20/30/40/50 and 30/30/30/30/30?

[Meridian]

I simply think this is not worth effort. Answer one question, what is difference in long run between 10/20/30/40/50 and 30/30/30/30/30?

If I understand correctly it was about difference between 30/40/50/60/70 and 30/30/30/30/30.
Which is a bit of a difference.
But I agree, it's just a nice to have, maybe after OpenXcom 5.0 is out

[Arthanor]

I simply think this is not worth effort. Answer one question, what is difference in long run between 10/20/30/40/50 and 30/30/30/30/30?

Expected number of landed hits:
1.5 for either configuration

Chance to fail to land a hit:
with 10/20/30/40/50: 0.9*0.8*0.7*0.6*0.5 = 0.1512 ~ 15%
with 30/30/30/30/30: 0.7^5 = 0.16807 ~ 17%

Chance to land all shots:
with 10/20/30/40/50: 0.1*0.2*0.3*0.4*0.5 = 0.0012 ~ 0.12%
with 30/30/30/30/30: 0.3^5 = 0.00243 ~ 0.24%

Alright, it's not that different, but it is different. The results will also vary for other numbers of hits, but those are more complicated and I'm at work. The point is, although both weapons have the same expected number of shots, one is more reliable to land a hit, while the other is more likely to land many hits. We tend to just consider averages (average outcome is the same, everything is the same), but there's more to distributions of numbers than that.

Now is it worth the extra coding? I can't tell because my coding for OXC has been minimal and it would be a lot of work for me (discover how to add new ruleset item properties, amongst other things). Using a simple quadratic formula like we have for damage scaling, but instead defining chance to hit as "F*(a + b*i + c*i^2)" where "i" is the shot number in the autoshot (so i = 2 means the 2nd shot) would enable one to define the effect of "walking your shot" (or alternatively, recoil messing up with your aim on rifles, thus decreasing accuracy as more shots are fired by defining negative parameters, or both) for autoshots.

[Yankes]

If I understand correctly it was about difference between 30/40/50/60/70 and 30/30/30/30/30.
Which is a bit of a difference.

Then this would be equal comparing two weapons that have 50 and 30 accuracy.
I ask about that have similar performance.

Expected number of landed hits:
1.5 for either configuration

Chance to fail to land a hit:
with 10/20/30/40/50: 0.9*0.8*0.7*0.6*0.5 = 0.1512 ~ 15%
with 30/30/30/30/30: 0.7^5 = 0.16807 ~ 17%

Chance to land all shots:
with 10/20/30/40/50: 0.1*0.2*0.3*0.4*0.5 = 0.0012 ~ 0.12%
with 30/30/30/30/30: 0.3^5 = 0.00243 ~ 0.24%

Alright, it's not that different, but it is different. The results will also vary for other numbers of hits, but those are more complicated and I'm at work. The point is, although both weapons have the same expected number of shots, one is more reliable to land a hit, while the other is more likely to land many hits. We tend to just consider averages (average outcome is the same, everything is the same), but there's more to distributions of numbers than that.

Fully agree, there are different but is this difference big enough to be something more that some kind of flavor?

Now is it worth the extra coding? I can't tell because my coding for OXC has been minimal and it would be a lot of work for me (discover how to add new ruleset item properties, amongst other things). Using a simple quadratic formula like we have for damage scaling, but instead defining chance to hit as "F*(a + b*i + c*i^2)" where "i" is the shot number in the autoshot (so i = 2 means the 2nd shot) would enable one to define the effect of "walking your shot" (or alternatively, recoil messing up with your aim on rifles, thus decreasing accuracy as more shots are fired by defining negative parameters, or both) for autoshots.

Implementing this is not problem, 90% code would be boilerplate needed to propagate valuer form ruleset to place where accuracy is calculated.
Its simply too shallow, there no gameplay there. Proper version of this should look like:
a) Your accuracy affect positive bonus per shoot (your good eye allow you fix your aim).
b) Your strength affect negative bonus per shoot (your steel grip don't move when you split hot lead).
c) using one hand for two hand weapon give you 0 accuracy after first shoot (Rambo exists only in moves).

But this would need lot more work.

[Arthanor]

Yes, I agree that the difference doesn't affect gameplay much, but over 10-20 shots (not uncommon in XPiratez), seeing the last few shots always hit instead of everything being scattered equally could be more satisfying. Also seeing rifles with multiple shots deteriorate as the autoshot goes would be funny. It does contribute to the experience even if it doesn't affect the stats much. Even better if somehow hit location of the previous bullet can be factored in, maybe reusing some code of the shotgun fix.

I also agree that a deeper implementation would be much more interesting. It would help further diversifying if the a, b, c coefficients I suggested could themselves depend on strength, firing accuracy and setup (flying makes recoil worse as you have nothing to stabilize against, kneeling makes it better as you are braced). But constant coefficients are a first (baby) step that already enables a little bit of something (and might be able to placate the request, as I have seen it multiple times already).

As I think that changes of distribution do affect player experience, I just wanted to chime in and support the idea. I fully understand why it isn't (and shouldn't) be a priority. But it shouldn't be entirely dismissed either. As Meridian said, for 5.0!