OpenXcom Forum
OpenXcom => Open Feedback => Topic started by: adam on December 20, 2019, 05:11:30 am
-
The accuracy model of this game is ludicrous. I just watched rookies shoot 60 degrees off-target for a distance that couldn't have been even 100 yards.
I don't really blame the devs; this is the way the original worked. It was stupid then too, but less noticeable, given all the limitations of DOS games.
Defending ourselves from aliens we can't rustle up rookies who can aim something like on-target? In a world with many billions of people?
The placement of ships is similarly stupid.
-
.
-
Vanilla will stay vanilla.
However, OpenXcom offers you a ton of modding possibilities and you can easily change your soldier's accuracy and the placement of ships to suit your desires.
-
Defending ourselves from aliens we can't rustle up rookies who can aim something like on-target?
You want too much. XCom has the best scientists in the world. In a year's time they came up with an interplanetary vessel armed with fusion weapons. Even if it is claimed that is just stolen tech from the aliens, they also came up with several laser weapons all on their own in about a month. Ronald Reagan spent years and billions of dollars with thousands of scientists on his "star wars program" and didn't even get a viable laser pistol design. XCom also has the best engineering and machine shop operation in the world. These guys built those light weight, accurate, no spare battery packs required laser weapons. But even more incredible, they built that fusion weapon armed interplanetary vessel in about a month all onsite mostly out of recycled junk. In comparision, Boeing which is a much bigger organization with a much bigger budget and much more experience designing and building aircraft couldn't get their software right on their 737 Max nor get their Starliner to the ISS.
As for the soldiers, well see the best I can figure it, XCom doesn't use trained soldiers. After all, if they started with special forces guys or the "best of the best of the best," these rookies wouldn't get over 3 times stronger over the course of a year. Xcom also doesn't seem to have any basic training capability either, instead preferring to do "on the job training" during missions. Unlike the mib guys, xcom never developed a neuralyzer. I think XCom drafts or conscripts into service any civilian that reports any ufos or little grey men in order to keep the whole alien thing secret. If a bunch of draftees get killed fighting aliens, well at least they can't spill the beans about flying saucers. If XCom is successful and the old surviving veterans years later tell stories about fighting big green men, secret government agents can discredit them by reminding the public these guys also claimed orange snake people terrorized their neighborhood in the late 90's. Similar to how in the movie "Independence Day" for years no one took the character Russel Casse, played by Randy Quaid, seriously about being abducted by aliens.
-
.
I don't think I've seen this before. That image is hilarious (and accurate ...well, accurate about being inaccurate lol).
-
The nice thing about openxcom's accuracy model is that a shot with a higher listed accuracy value will not veer off target by such a large angle. This means that the closer your target is, the less accuracy you need to reliably hit it. This was not so in the original game, where the accuracy value only gave you the chance of getting a perfect hit, while all non-perfect hits had randomized stormtrooper accuracy and only reliably hit adjacent targets. Even 1 square distance was enough to cause most "misses" to actually miss.
edit: openxcom extended has the improved accuracy model
-
This was not so in the original game, where the accuracy value only gave you the chance of getting a perfect hit, while all non-perfect hits had randomized stormtrooper accuracy and only reliably hit adjacent targets. Even 1 square distance was enough to cause most "misses" to actually miss.
As far as I know, this is not true.
Original worked the same as OpenXcom.
-
the accuracy model in openxcom is 1:1 with the original, we went to painstaking efforts to ensure that.
-
the accuracy model in openxcom is 1:1 with the original, we went to painstaking efforts to ensure that.
My preliminary testing in the GOG version of the game would seem to confirm that, however that is most assuredly NOT the experience I had back in the day playing the collector's edition. In openxcom, a shot with 75% accuracy will never veer 40 degrees off course but it routinely did in the old game. Any high accuracy shot taken several times would show mostly precision hits along exactly the same shot path (or precision misses, if it's a bad path), while all of the shots that aren't a perfect hit would just be randomly aimed in approximately the same direction as the target. It seemed as though having the target be closer caused the maximum shot angle to increase, and I've seen a point-blank shot veer off by as much as 90 degrees from its target in the old game, probably the collector's edition. It was so much of a problem that I ran several tests and confirmed that low-accuracy shots taken with one space distance from the target will miss more often than they hit. I even discovered that the angle to the target can drastically alter the chance to hit them. A target standing orthogonally adjacent to the shooter has something like a 90%+ chance to get hit with a missing shot, while a target diagonally adjacent and one space lower than the shooter will get hit by around 10% of misses or less.
Maybe it's OXCE, but a shooter listing over 50% accuracy will always have a narrow shot angle with degrees in the single digits. Firing at a target ten squares away, they will rarely miss by more than a square. And in my preliminary testing on the GOG version of X-Com (just now), that appears to also be the case. If that had always been the case, then where did all of these X-Com accuracy memes come from? My very significant amount of experience firing weapons in open xcom extended not only contrasts heavily with my very significant amount of experience firing weapons in the old games, I've also never at any point in time playing OXCE had any provocation to despise the OXCE accuracy system. (Though when I played OXC, I recall having a ton of problems with it.)
Actually I remember now. OXCE fixes accuracy. It was when I switched to OXCE that I stopped having accuracy problems. Previously, the accuracy problems were so bad that it was virtually impossible to mod my way around them. But with OXCE accuracy, it works so well I almost don't need to make any changes.
-
Accuracy in OXC and OXCE is identical. 100% sure.
-
There was this thing discussed hundreds years ago about a difference between vanilla and OXC targeting:
[...]
Vanilla got only 3x4 grid - just 12 target points, so in most of cases the target voxel was at least 1 voxel away from the obstacle, and AIM was hitting the target. In OpenXcom we have a grid 3x10, with really precise vertical, which OFTEN choses the voxel at the edge of obstacle, which instantly lessen chances by 50% for AIM.
Dunno if it was ever changed.
-
There was this thing discussed hundreds years ago about a difference between vanilla and OXC targeting:
(quote from 2015)
Dunno if it was ever changed.
it was, repeatedly, volutar ensured that it was identical.
-
Thanks for the clarification and the commitment.
-
There is one difference oxc definitely has over the original--and it might just be a visual glitch. But shots regularly travel through the target and keep going, a miss even though it looks like a direct hit. It is difficult to demonstrate it on video, it just happens randomly every now and then. But in playing a mod which gives weapons much lower accuracy than I am used to, I noticed it happening far more often. I was making point-blank auto shots (1-4 squares between shooter and target) with listed accuracy below 20%, and seeing multiple shots in a row go center of the target and keep going, finally impacting on something behind the target. I NEVER saw that in the original.
-
what you see on-screen and what the game uses behind the scenes are two very different things.
you see a 2D sprite-based representation/interpretation of a 3D voxel-based world.
if anything, it's a visual difference only.
take my word for it, the trajectory calculation is as identical to the original as possible.
-
...and seeing multiple shots in a row go center of the target and keep going, finally impacting on something behind the target. I NEVER saw that in the original.
Here, attached video from the original... a shot goes right through the floater's right arm and impacts on the wall behind him (at 1:30).
Took just a few minutes to reproduce.
As Warboy said one post above, 2D sprite and 3D model are just different, no magic involved.
Shots going (only visually) right through the legs and arms are quite common.
As for going straight through the dead center of the unit... that should not happen in vanilla, but can easily happen in mods.
But even in vanilla, it can (only visually) LOOK LIKE it goes straight through the center of the unit... depending on the angle of the camera and angle of the shot.
For example, a soldier shooting directly upwards (i.e. from screen bottom to screen top) and a (missing) shot going just ever so slightly above the target's head in 3D space... will definitely look like a headshot or chestshot that goes straight through. Such situations are less common, but happen every now and then.
-
As for going straight through the dead center of the unit... that should not happen in vanilla, but can easily happen in mods.
It happens in openxcom without mods. It happens a lot in some cases. I haven't figured out what makes it happen more but I suspect lower firing accuracy is at play, given that the X-Com Classic Remastered mod (which doesn't change the hitboxes) uses a lot of low firing accuracy and has it happen a lot, whereas my own Reaver's Faithful mod has a lot of high firing accuracy (also doesn't change hitboxes) and doesn't get the issue happening as much as it does in just plain unmodded.
-
It happens in openxcom without mods. It happens a lot in some cases.
Well, if it happens a lot, make a video of it, post it here and we will explain what is going on.
-
the accuracy model in openxcom is 1:1 with the original, we went to painstaking efforts to ensure that.
Oh, I concede it. I'm alternately infuriated and greatly amused.
Someone should make a game like this, but a little less... insane.
-
I posted another topic on this, and then realized I'd already said it. It keep blowing my mind.
If our soldiers aimed like this, we would have lost EVERY war.
-
Feel free to headcannon it.
For me, your men aren't soldiers at all. They are government agents, more like the FBI or the KGB. While trained, they cannot compare with actual military personnel.
-
Another example: Base raid, I walk past an "arrowhole" window in a hallway and the alien shoots me. I reload, replay it a bunch of times, alien kills me maybe 3 out of 5. But, if at the point where he kills me I wheel and shoot -- well, I can't. "No line of site." The path that works for the alien apparently doesn't work in the opposite direction.
It's stupid and aggravating.
-
Another example: Base raid, I walk past an "arrowhole" window in a hallway and the alien shoots me. I reload, replay it a bunch of times, alien kills me maybe 3 out of 5. But, if at the point where he kills me I wheel and shoot -- well, I can't. "No line of site." The path that works for the alien apparently doesn't work in the opposite direction.
It's stupid and aggravating.
Additionally to the above, line of sight is calculated (roughly) from the unit's eye level, while the shooting line is calculated (roughly) from the weapon's level. Naturally there are many cases where an alien can see and shoot you, but you can only see them, or only shoot them - especially when the alien is of a different height (like a Sectoid) or standing on something.
These are some basic tenets of X-Com mechanics, and I really don't think anyone will be interested in changing them. People have asked before, but it would break the compatibility with vanilla game and is therefore out of the question.
Sorry, but no amount of complaining, no matter how valid, is going to do anything about it. The only sensible response is "it's an open project, if you want it different, make a clone and code it yourself". I am not trying to be an asshole, this is simply how it is. [Devs: If I'm wrong, please correct me.]
-
Sorry, but no amount of complaining, no matter how valid, is going to do anything about it. The only sensible response is "it's an open project, if you want it different, make a clone and code it yourself". I am not trying to be an asshole, this is simply how it is. [Devs: If I'm wrong, please correct me.]
You are 110% correct.
-
on this topic, is there an understanding of how close range increases the accuracy of fire in xcom? ufopaedia has an excellent article demonstrating it happens (https://www.ufopaedia.org/index.php/Firing_Accuracy_Testing), but there does not seem to be a general formula for this topic unless it has been discovered elsewhere.
A quick glance at the figures on ufopaedia would suggest approximately a x1.79 modifier at range 2, x~1.483 at range 3, and x~1.236 at range 5. But of course these crude estimates are probably wrong.
-
on this topic, is there an understanding of how close range increases the accuracy of fire in xcom?
https://www.ufopaedia.org/index.php/File:FAT-FFA45.png (https://www.ufopaedia.org/index.php/File:FAT-FFA45.png)
The graph in the article seems to show that shot groups get slightly tighter the further away the target is. The model spread line is a straight diagonal dissecting the graph, indicating a firing angle of 90ยบ total, regardless of distance to target. But the actual spread begins slightly higher than the model, but gets slightly lower with greater distance on a curve that appears if the chart went out to 60, they would deviate greatly.
I'd say this concurs with my own experience in the original game. Many times I noticed that getting closer to the target didn't help much because the shot angle just got wider, that is until they got adjacent with the target in which case suddenly all shots started hitting.
I've ran a considerable amount of testing attempting to discover any such angle change using both OXCE and UFOExtender accuracy, and generally speaking, the weapon always seemed to fire with exactly the same firing angle at any range. It would only change when hit chance reduction from being outside of visual range took effect. I was able to create scenarios in which the angle changed, but only with bizarre setups that result in listed firing accuracy going through a drastic change over a short distance, ie. with a shot range of 15 and a dropoff of 5. Maybe UFOExtender accuracy is somehow mediating and making the result appear to maintain the shot group, but one way or another it seems to just work, where before it didn't.
-
I'd say this concurs with my own experience in the original game. Many times I noticed that getting closer to the target didn't help much because the shot angle just got wider, that is until they got adjacent with the target in which case suddenly all shots started hitting.
yeah, I'm not eager to run testing but I want to get an understanding of the aproximate multiplier. Right now I'm trying to figure out the true 1-fire action-kill effectiveness of various weapons vs various aliens (IE: multiplying accuracy * % of the time that the weapon has sufficient damage kill an alien), and realize that without the proper inputs for close-range fire I can't accurately model the true effectiveness of auto fire. Cursory attempts to trial weapon accuracy suggests noticable improvement at range 3, but I really need to get a set of numbers
-
Oh, I wasn't expecting anyone would fix it; I've seen enough threads here to have the vibe down. But the title of the topic is "open feedback". It's feedback. As with line of sight, I get how the game works -- it's stupid and irritating, but hey, that is how the original worked, too. It was less irritating when DOS was the program loader and it took trickery to get more than 640k. In the modern era...
But sure, your goal is to replicate the original (except you don't, you add all sorts of bells and whistles). It's a great achievement, in that sense.
There are a million of these things. My soldier (er, "agent", if that makes you feel better) steps forward and suddenly the muton pops out of nothingness. It made sense in the early 1990s. Now... it seems bizarre.
Maybe I'm just too old. I've played a lot of X-COM over the years, and openxcom has been part of that, but... I just can't suspend disbelief like I once could.
-
There are a million of these things. My soldier (er, "agent", if that makes you feel better) steps forward and suddenly the muton pops out of nothingness. It made sense in the early 1990s. Now... it seems bizarre.
I didn't think it made sense back then. But hey, that's the beauty of modding: I get to fix the game's problems the way I think it should be!
-
I just meant it made sense given the technological limitations... there was no great way to show that you could only kinda/sorta see them.