|
Post by willagaroon on Dec 6, 2010 15:03:17 GMT -5
I'm starting a new thread because I thoroughly hijacked that last one. To recap, I'm looking for anyone to contribute knowledge to increase the accuracy of this simulator. Need things like how center speed might work and if anyone disagrees with the assertions I list below. Okay then, here is a simulator for all the AR's. This is based on rolling a single random number between the two points on an axis. If the two points fall on the same side (L96A1 in the up direction for example) it rolls a random number between the two numbers (between 80 and 100 for the L96A1). This does NOT take into account center speed because I'm not sure how that works. If someone has an idea please let me know. This is really just for fun since without knowing exactly how all the numbers are handled I can't do this 100% accurate. It does do a good job of showing the randomness though. Oh, please don't link to this outside of this forum otherwise I'll have to pull it down to save the bandwidth. www.igameguides.com/util/recoil.aspx?rounds=30And you can specify the number of rounds to simulate, from 1 to 400.
|
|
|
Post by KingVaroon on Dec 6, 2010 15:10:30 GMT -5
Does not seem correct. It shows most of the guns kicking downwards more rather than up.. and it shows the G11 to actually have recoil
|
|
novem
True Bro
Posts: 193
|
Post by novem on Dec 6, 2010 15:13:31 GMT -5
Because it has, specifically the same as famas does, but it's 3 burst rifle and has better centerspeed, therefore better overall accuracy.
|
|
|
Post by willagaroon on Dec 6, 2010 15:17:26 GMT -5
novem, you mentioned in the old thread the problem with FAMAS kick down when it should kick up. Well as the other guy mentioned in this thread it was a common problem for ALL weapons on that chart.
I screwed up the math doing a change but it's fixed now, they kick mostly up now as they should (was adding when I should be subtracting).
|
|
|
Post by KingVaroon on Dec 6, 2010 15:17:41 GMT -5
You missed my point. This simulator does not work.
edit: I see you corrected it. Much better. doesnt seem completely accurate though.
|
|
|
Post by willagaroon on Dec 6, 2010 15:23:34 GMT -5
You missed my point. This simulator does not work. The down kick versus up kick has been fixed, they mostly kick up now, it was a math problem as I said. As for it not being accurate, I don't have the actual working math behind recoil and center speed to make this 100% accurate. I have to go by best guesses based on the information out there. I've already said that these sims do NOT take center speed into account because I haven't seen anyone suggest a mathematical way to do that. I'm not doing this to recreate the guns 100% I'm just doing this because it's interesting and to attempt to get as close to a working sim as possible for entertainment and discussion purposes. edit: whew, so you see it's better now...i'll leave all this above for others to understand what this is and what this isn't
|
|
|
Post by KingVaroon on Dec 6, 2010 15:27:13 GMT -5
Great post btw. Once someone is able to explain how exactly centerspeed works, this will be really helpful.
|
|
aranshada
True Bro
But woe betide thine enemies, for thee hath created thine Bullet Hose.
Posts: 231
|
Post by aranshada on Dec 6, 2010 15:36:45 GMT -5
Again, setting it to 400 rounds produces a nice little box. =)
Perhaps the CenterSpeed is actually a measure of degrees of recovery per second where recoil kick is measured as degrees of kick.
Of course, no weapon kicks 60 degrees, so perhaps those are the x10 versions of them, whereas the max kick for the Commando would actually be 6 degrees. CenterSpeed could then be taken roughly as 150 degrees of recovery per second which would make a 6 degree kick from a Commando take 40 milliseconds to recover fully from: 150 degrees per second / 1000 milliseconds = .15 degrees recovery per millisecond; 6 degrees / .15 recovery per millisecond = 40 milliseconds. Keep in mind, that equation took me entirely too long to figure out (or at least get a answer that seemed reasonable)... it's been too long since I've had any sort of math course, and I am quite sleepy.
Anyway, if the above calculations pan out, then... nevermind, I just checked the chart and the firetime for the Commando is .08 which would mean that my calculations would have the weapon recover from a full kick before the next shot ever fired which would result in a perfectly accurate weapon...
I would advise that you disregard this post, but I'll leave it here for future reference and to possibly spark other ideas. Good day, sir.
|
|
|
Post by esp211 on Dec 6, 2010 15:56:52 GMT -5
I love threads like this.
But how many times does anyone spray their entire clip at a non-moving target? Personally I compensate for recoil by moving my right stick down while spraying but only when the target is moving and relatively up close. Most of the time I burst fire but still compensate if I don't kill right away. This is why I think it always comes down to personal preference because all guns have advantages/disadvantages based on the player's style.
|
|
aranshada
True Bro
But woe betide thine enemies, for thee hath created thine Bullet Hose.
Posts: 231
|
Post by aranshada on Dec 6, 2010 16:02:04 GMT -5
They don't spray the entire clip, but they might spray 5 shots. If you can come up with an average of how those 5 shots will land, then you'll know exactly how to adjust that FAMLAZER by aiming slightly left to counter it's slightly right recoil, or you'll know that there's nothing you can do with the commando and you should just keep spraying until it randomly puts its own bullets back on target.
It's more for education than anything. And according to the OP, partly because he's bored and wanted to see if he could do it.
|
|
|
Post by SheWolf on Dec 6, 2010 16:05:05 GMT -5
i'm still thinking we should consider that 2 variables are made for each axis, not just one.
|
|
|
Post by willagaroon on Dec 6, 2010 16:08:03 GMT -5
i'm still thinking we should consider that 2 variables are made for each axis, not just one. I actually started this simulator going that direction (roll twice per axis) and while it does change things, it doesn't change them drastically over a large sample size. I'm not giving up that it could be done another way, just throwing that out there.
|
|
|
Post by SheWolf on Dec 6, 2010 16:18:05 GMT -5
i think it would make more sense for 2 reasons:
- you get a neat gaussian distribution along each axis, wich would be preferably gameplaywise i think - certain weapons that have two up-arrows in the chart indicate that there are two numbers at work
great work though, really neately done...and it shows once and for all that the famas, the french masterpiece, is superior to this austrian trash they call the aug^^
|
|
|
Post by willagaroon on Dec 6, 2010 16:31:50 GMT -5
i think it would make more sense for 2 reasons: - you get a neat gaussian distribution along each axis, wich would be preferably gameplaywise i think - certain weapons that have two up-arrows in the chart indicate that there are two numbers at work Just so I'm clear on what you are saying. And before someone sees this and says it's the wrong way, I'm just hearing him out and I'm NOT changing the sim at this time. For example the M16 with 35-50-30-30 (L-U-R-D) would be calculated as follows: L = Random # between -35 and 0 R = Random # between 0 and 30 x-axis offset = L + R ...repeat for U/D and y-axis That makes sense but what about the one's with two numbers in one direction? It can't be the same formula otherwise it would kick way too much in that direction. What do you think?
|
|
|
Post by SheWolf on Dec 6, 2010 16:41:17 GMT -5
yeah, you nailed it, that's what i meant. (allthough i believe the m16 has 35 left 35 right, but you're point is valid)
the ones with the the two numbers in one direction (or negative number in the native direction, does the same thing) are mostly shotguns and sniper rifles, and those things do kick a lot... oh, and the m60. and some of the handguns..
|
|
novem
True Bro
Posts: 193
|
Post by novem on Dec 6, 2010 17:19:26 GMT -5
L = Random # between -35 and 0 R = Random # between 0 and 30 x-axis offset = L + R ...repeat for U/D and y-axis That makes sense but what about the one's with two numbers in one direction? It can't be the same formula otherwise it would kick way too much in that direction. What do you think? If you have for example <lower number,higher number> i think Random # between them. But I understand that you can use only random between 0 and some other number, so the formula is: offset= lower number + (Random # between 0 and (higher number - lower number)) It works for lower than 0 numbers too, for example <-10,60>, u have -10 + (random between 0 and 70), so u can roll every number from [-10,60]. It's still 1 roll and i really don't understand why u want to have 2 rolls. Because there are 2 numbers? Yes, but the point of roll is to generate 1 number between them.
|
|
|
Post by imrlybord7 on Dec 6, 2010 17:38:12 GMT -5
The fact that guns with two positive numbers are guaranteed to kick in that direction basically proves that there are only two numbers in play for viewkick, not four. While they could have a system where if the gun had a positive and negative it would pick two to average and if had two positives it would just pick one, they wouldn't. Absotively, posolutely wouldn't.
|
|
|
Post by SheWolf on Dec 6, 2010 18:21:57 GMT -5
au contrair, i believe this weapons prove that there are 4 numbers. otherwise why give the olympia 85 up AND 95 up?
i think you misunderstood the 2-random-numbers-per-axis-theory.
|
|
|
Post by imrlybord7 on Dec 6, 2010 18:58:19 GMT -5
... Are you really asking that? It guarantees that it will kick up anywhere from 85 to 95 units of CoDkick regardless of what theory you are operating under. And I really don't see why they would pick two numbers. Den himself has stated that averaging together the min and max gets you the average kick value for that axis. What would be the point of picking two numbers instead of one? That would make it almost impossible for a gun to kick toward its less likely direction.
Example
A gun with -10 to 60 for its x axis would pick one negative value from -10 to -1 and one positive value from 1 to 60 (this punches another big hole in the 2 per axis theory - what the Foxtrot happens to zero?). The chance of the negative number having a greater absolute value than the positive number is minuscule, whereas the 1 per axis theory gives you a 10/71 chance of kicking to the left, which is actually pretty significant.
EDIT: Since I'm bored and feel like doing some of the math, the minuscule probability of kicking left under the 2 per axis theory comes out to be 9/60 (number of positive values that can be picked ans still have it kick left) minus the probability of the positive value having a greater absolute value than the negative value anyway. If anyone can calculate that probability then we can find the actual probability of kicking left under such circumstances, which I assume would end up being rather small.
|
|
|
Post by SheWolf on Dec 6, 2010 19:23:33 GMT -5
that's my whole point: den writes on the page that a 10-60 gun will allways kick to the right. thats why i doubt that the number is picked between -10 and 60.
concerning what would happen to zero: the sights would not move along that axis. but zero can also happen in the 1 per axis theory.
|
|
|
Post by imrlybord7 on Dec 6, 2010 19:30:51 GMT -5
No. Zero can ONLY happen in the 1 per axis theory, which is my argument. Unless they assign 0 to be positive, which would just be weird and another reason that they wouldn't go with that. And obviously a 10 to 60 gun will always kick right because IT WOULD HAVE NO NEGATIVE VALUES. A -10 to 60 gun, however, could kick left.
|
|
|
Post by willagaroon on Dec 6, 2010 19:38:10 GMT -5
No. Zero can ONLY happen in the 1 per axis theory, which is my argument. Unless they assign 0 to be positive, which would just be weird and another reason that they wouldn't go with that. And obviously a 10 to 60 gun will always kick right because IT WOULD HAVE NO NEGATIVE VALUES. A -10 to 60 gun, however, could kick left. Again, I'm not picking sides with this because I don't know the answer, but a 0 CAN happen in the two roll theory since it's two random numbers from -10 to 0 and 0 to 60. At least in the way that I think about it (-10 min 0 max, 0 min 60 max). Of course for it to land at 0 it would have to randomly roll 0 TWICE which isn't as unlikely as you might think at first since the -10 to 0 roll will randomly be 0 one out of eleven times. Rereading I'm still not thinking you are getting the two roll theory. It is -10 to 0 and 0 to 60 which means that it absolutely CAN be a negative number.
|
|
|
Post by imrlybord7 on Dec 6, 2010 19:46:01 GMT -5
EDIT: I failz at maths.
(9/60)*(9/10)*(1/2)=81/1200.
That's REALLY close to the actual probability of our theoretical gun kicking left assuming the 2 per axis theory, which is just under half of the probability of it kicking left assuming the 1 per axis theory. I don't see why they would make it that unlikely for a -10 to 60 gun to kick left.
|
|
|
Post by SheWolf on Dec 6, 2010 20:02:30 GMT -5
No. Zero can ONLY happen in the 1 per axis theory, which is my argument. Unless they assign 0 to be positive, which would just be weird and another reason that they wouldn't go with that. And obviously a 10 to 60 gun will always kick right because IT WOULD HAVE NO NEGATIVE VALUES. A -10 to 60 gun, however, could kick left. Again, I'm not picking sides with this because I don't know the answer, but a 0 CAN happen in the two roll theory since it's two random numbers from -10 to 0 and 0 to 60. At least in the way that I think about it (-10 min 0 max, 0 min 60 max). Of course for it to land at 0 it would have to randomly roll 0 TWICE which isn't as unlikely as you might think at first since the -10 to 0 roll will randomly be 0 one out of eleven times. asside from the 0 vs 0-scenario, you can also have a -1 vs 1, -2 vs 2.... -10 vs 10.
|
|
kalar
True Bro
Attack! Attack! Attack!
Posts: 451
|
Post by kalar on Dec 6, 2010 20:08:04 GMT -5
Cool. Can we get some graphs for the non-AR weapons? Centerspeed and the mathematical specifics are way over my head, but best of luck to you.
|
|
|
Post by imrlybord7 on Dec 6, 2010 20:17:49 GMT -5
For easier comparison the 1 per axis theory gives a 14.08% chance of the theoretical gun kicking left while the 2 per axis theory gives around a 6.75% chance.
|
|
|
Post by caboose on Dec 6, 2010 20:20:51 GMT -5
So let me see if I got this straight:
Currently, without speed center implementation, this generator only shows all the places the second bullet can go? (The first always goes on target)
EDIT: Actually, it doesn't show where the bullet goes at all, does it? Doesn't it just show how much recoil you can potentially get from a single shot? (But then center speed negates some of that)
|
|
|
Post by willagaroon on Dec 6, 2010 20:47:37 GMT -5
This does show where the bullets go in a world without center speed and based on our interpretation of the recoil calculations.
Add in center speed and what should happen is guns with higher (faster) center speed should "tighten" up the grouping.
For now these charts are really just something interesting to look at and can show general recoil patterns (nothing you couldn't get from simply looking at the recoil numbers, but hey, pictures are more fun to look at).
If anyone can provide insight into center speed calculations I can certainly add that to these charts.
|
|
|
Post by imrlybord7 on Dec 6, 2010 21:03:45 GMT -5
While you might think that ignoring center speed means you can ignore ROF, you would be mistaken. If a gun fires fast enough the next shot will come out before the center of the screen has gotten all the way to the point of kick, and if it fires slow enough the next shot will come out after the center speed has started the recovery. It is also possible for a gun to fire at just the right speed where it fires at the peak of its kick (Den once told me that this causes the MP5 from CoD4 to exhibit more recoil than the numbers would lead one to believe). I also know for a fact that there is no "residual" kick. Once the next shot comes out, regardless of whether or not the previous shot made it all the way to its point of kick, that shot's view kick value is the only one in play. That's why sometimes a higher ROF can lead to lower recoil.
|
|
aequinox
True Bro
hakuna matata
Posts: 366
|
Post by aequinox on Dec 6, 2010 23:14:32 GMT -5
They show the first shot on target, then it rolls a second shot n-1 times (with n being variable in the URL) and shows that. This becomes obvious as when n tends to a very large number, a rectangle is formed as opposed to a continuous line of recoil.
|
|