I understand and appreciate the recent adjustment to the UAC/5...
That said, I think the probability of jamming on the first shot is a little extreme, I have two potential suggestions to balance between the original iteration and the recent hot-fix:
1.) Randomize the jam percentage (between 5% to 25%). - This would allow the jams to "feel" random. Right now I completely anticipate the first or second double-tap to jam... guaranteed.
2.) Allow the first double tap to go unpunished. - The propensity of the very first double-tap to jam is quite high (I estimate it's over 50%) essentially unfairly punishing the user for using the weapon. User should at the very least be able to get the first shot off before being potentially penalized.
Please understand I am in support of the recent hot-fix but believe it's a little too ridged in it's application.
Thanks.


[Suggestion] Uac/5 Jamming
Started by DaZur, Nov 23 2012 06:50 AM
5 replies to this topic
#1
Posted 23 November 2012 - 06:50 AM
#2
Posted 23 November 2012 - 08:13 AM
I think you mis-understand randomness.
The jamming percentage chance is currently 25%.
Each time you double-tap, there is a 25% chance of jamming. Every time.
You can't randomise the random factor further. Well, you can, but to what end? If you want to randomise "between 5 and 25%", all you're doing is randomising it to 15% instead of 25%
No, the first double-tap isn't 50%, it's 25%, just like every other time you double-tap. If you let people get away with one double-tap, then that's (marginally, but still) exploitable.
The jamming percentage chance is currently 25%.
Each time you double-tap, there is a 25% chance of jamming. Every time.
You can't randomise the random factor further. Well, you can, but to what end? If you want to randomise "between 5 and 25%", all you're doing is randomising it to 15% instead of 25%
No, the first double-tap isn't 50%, it's 25%, just like every other time you double-tap. If you let people get away with one double-tap, then that's (marginally, but still) exploitable.
#3
Posted 23 November 2012 - 08:17 AM
Too bad the UAC5 is now jamming on non double tap's as well. I have had it jam on my first shot fired.
#4
Posted 23 November 2012 - 09:05 AM
Knyght, on 23 November 2012 - 08:13 AM, said:
I think you mis-understand randomness.
The jamming percentage chance is currently 25%.
Each time you double-tap, there is a 25% chance of jamming. Every time.
You can't randomize the random factor further. Well, you can, but to what end? If you want to randomise "between 5 and 25%", all you're doing is randomizing it to 15% instead of 25%
No, the first double-tap isn't 50%, it's 25%, just like every other time you double-tap. If you let people get away with one double-tap, then that's (marginally, but still) exploitable.
The jamming percentage chance is currently 25%.
Each time you double-tap, there is a 25% chance of jamming. Every time.
You can't randomize the random factor further. Well, you can, but to what end? If you want to randomise "between 5 and 25%", all you're doing is randomizing it to 15% instead of 25%
No, the first double-tap isn't 50%, it's 25%, just like every other time you double-tap. If you let people get away with one double-tap, then that's (marginally, but still) exploitable.
LOL! I think you misunderstand...

A 25% chance of jamming is not random. it's a fixed off-set... Which means there is a constant variable being 1-in-4 chance.
My postulation is that variable be randomized further.. between 1-in-10 to 1-in-4 and all subsequent variables in between.

as it stands there is a 1-in-4 chance of your AC jamming on the first shot. That said, there is a (un-substantiated claim) higher probability of that first-shot jamming than not... Which leads me to believe it's not as random as one would think. (Read: weighted probability).
My suggestion creates a wider range of potential variables, thus creating a feeling of randomness while still limiting the min/max of the penalty.
#5
Posted 23 November 2012 - 09:28 AM
until I see proof of that with math to back it up Im going to say the 'jams on 1st shot every time' to be a perception bias.
#6
Posted 23 November 2012 - 09:49 AM
Sprouticus, on 23 November 2012 - 09:28 AM, said:
until I see proof of that with math to back it up Im going to say the 'jams on 1st shot every time' to be a perception bias.
Never said they jam 100% on first shot... Perception however is that the results appear to favor the likelihood of jamming more than not.

3 user(s) are reading this topic
0 members, 3 guests, 0 anonymous users