Rifleman Archive
Thread: The range mod posts in Corre Forums
I don't think it's accurate to say that giving a particular class significant dominance at a given range will make the game 'boring' or break balance, necessarily. The whole point of tactics is to put one's self in a position where one has an advantage. If pistols are better up close and rifles are better further away, then tactics are involved in arranging things such that you're fighting at your ideal range. You could certainly argue that a long-range advantage is superior to having a short-ranged advantage for other reasons (like radar making it difficult to 'sneak up' on people) but it's not inherently 'boring'. Having burst run be usable any time, but have a continuous HAM drain (as an example) would certainly make such a system quite viable.
I don't quite understand what the distinction between tests and events that you are making here. If I get 139 people to give me one sample, I don't see how I have any more (or less) information than if one person gave me 139 samples.
To answer the original post, the special mod to-hit (at least for BodyShot1) is additive, not multiplicative. The value I got is +20% (Noules got +25%). My char have battle fatigue, so that may be the reason for my lower value.
So what does this mean? First, +25% is equivalent to a +50 accuracy modifier. That's equivalent to 5 to 7 advancement box (some boxes are only +5 accuracy with the majority being +10, so the # of boxes advanced depends on your current skill level). Put it another way, it's equivalent to gaining 1,250,000 pistol xp. If dev don't see this as unbalancing, they got the head up the AAAaaa, erh, Arm. The fix for this is to reduce the mod to +5% to +10% (equivalent to 1 or 2 boxes gain).
The second problem is the special mod to-hit is additive, so it's just as effective at max range than at close range. This doesn't feel right. In RL target shooting, even if one aim for the torso, one will hit less often at long range than at close range. The fix is to have the special mod to-hit be affected by range. Either by using the current weapon range modifier or else just a built-in range modifier. Using the current +25%, let's say at ideal, it's +25%, at long range +5% or +0%.
Even though the specific example is pistol base, dev need to take a look at all special move to-hit modifier to determine if they're out of whacked.
Note 1:
How I came up with these number: I did 21 tests from 1m to 64m, 3m interval. Each test consists of 100 auto-attack shots. Then I re-ran the tests, this time using BodyShot1 instead of auto-attack. Since everything else is the same except the attack style, all the differences are due to the attack style. I plotted the result of both test and use the linear best-fit for both and then compare against the weapon range modifier. Both best-fit line have a r-square value of 0.998 (ie <1% precision). The slope of both lines are half that of the weapon range modifier, implying 1) the to-hit % is half the range modifier value (collaborating Noules that to-hit % = range modifier/2) and 2) the special to-hit modifier is additive since the slope is the same for auto vs special. The value of the addition I got is 20% with a covariant # indicating <1% precision. This is different than Noules 25%, hypothesis is Battle Fatigue reduce the to-hit gain.
Note 2:
The sqrt(n/2) come from the Central Limit Theorem which states that histogram of a large number of random samples of a population with a known distribution is gaussian even if the population is not gaussian. For example, let say the height of the people in Flatland is a inverse parabolic between 5.0 feet and 6.0 feet. If you take 100 random people and determine the average height and then use the sqrt(n/2) to determine the standard deviation, you'll be way way way off. You would say the average height is 5'6" with a stand dev of only 7% (eg 68% of the population are between 5'5" and 5'7") and you'll be way way way off since the true distribution is parabolic, not gaussian. What you need to do is take 100 samples, each sample sample 100 random people. Plotting the histogram of the 100 sample, you'll have an average height of 5'6" with a stand dev of 18%, eg 68% of the population is between 5'2" and 5'10") and your precision of this data is 7%, eg if you're off, it'll be between 5'1.7" or 5'10.3")
For the measurement of a to-hit, it's simply the weighted coin problem. Say the probability of head coming up is 68%. If I flip 139 coins, how precise is my data? Use the gaussian curve to determine that. And, yes, for the spread, I used 2x stand dev, not stand dev since stand dev mean I'm wrong 31% of the time. 2x mean I'm only wrong 3% of the time. For 139 coins with a probability of 68% hit, 2x stand dev is 8% and the spread is 16%. Meaning, if 10 other players perform the same test of shooting 139 times, 9 out 10 should have the to-hit % within 16% of each other.
The better way to determine to-hit is not to increase the # of tries, but to vary some factor and use the linear best-fit. For sample, take 10 guys, each with different accuracy modifier. Have each shoot 139 times. Then plot and correlate the slope against the accuracy modifier. The r-squared and the covariant will tell how tightly related they are; eg it'll tell if if the to-hit% is really just Accuracy Modifier/some integer or not.
Since i don't have 10 char nor do I want to create 10 char, I can't perform this analysis.
Based on the fact that the weapon range modifier is 2x to-hit modifier, it's reasonable to assume the same apply to all other modifier (eg as you assert, to-hit is accuracy skill / 2).
Based on the fact that special mod is additive and not multiplicative, it's reasonable to assume that all modifier are additive, thus collaborating your equation of to-hit = base + (accuracy skill/2) + (weapon range mod/2) + (shooter posture/2) + (target posture/2) + (special mod/2).
Ah, I see. The level of confidence was what I was missing.
As a note, I don't think the shooter posture is divided by two. I determined the value of the postures by watching the accuracy meter next to the HAM bar (which should be what the game 'thinks' the values is) using /aim at times to 'boost' the accuracy meter as needed. Going prone as the shooter increased the accuracy meter by 30, and my guess is that this increased the to-hit by 30% (the problem is that I seem to have come up against the accuracy cap in that particular set - the miss rate drops from ~30% to ~5%). The crouched shooter set changes the accuracy by about 12%, and taken together leads me to believe that shooter posture changes are applied directly without the reduction.
Looking back at my data, it doesn't seem too unreasonable for BS1 to be +40 to-hit rather than +50, although +50 does fit better (andmost of the tested pistol specials ended up around that range).I'm not sure how much BF your character had - if the drop is really due to BF, it's probably a step-function. I think the wound healing reduction due to BF occurs around 350, so if your BF was much lower than that, BF can probably be ruled out. It seems a bit odd that it'd only apply to specials accuracy, as well.
What it's saying the darn to-hit formula is jacked and should be fixed.
Make the special mod to-hit be affected by range
Lower the special mod to-hit.
Increase the max weapon range modifier
Make the range modifier be multiplicative.
But please please please don't introduce range cap and/or special range cap.
I'm a pistoleer and I agree that pistol is too good on the move at max range using special. Just fix that. Don't fight for range cap. It'll make the game fairly boring (if I'm prevented from doing special I worked hard to acquire, the game is boring).
Actually you get no speed reduction at all unless you are a Master Rifleman. Let's take a look at the numbers.
For instance,let's sayyou use a 115-333 7.5 T21 rifle, which a Rifleman posted as a "typical" post-nerf T21 in a post recently. As a non-Master Rifleman (4-4-4-4, everything but the Master box) you fire this T21 at the rate of once per5.4 seconds using the now familiar speed equation:
(WeaponSpeed * SpecialDelayMod) - (WeaponSpeed * SpecialDelayMod) * (ProfessionSpeed/100)
And if result < 1.0s , result = 1.0 (speed cap)
Because you're not firing faster than once per three seconds anyway, the "nerf" doesn't affect the 4-4-4-4 Rifleman at all. It _does_ effect the Master Rifleman, however. Currently, the Master is firing this rifle once every 1.8 seconds! Right now most of the benefits of Rifleman are in that one box.
Just realize what's going on here. The result is a zero reduction for non-Masters and a 40% DPS reduction for Masters. I think the solution is not to ask to kite other classes in exchange or to ask for the Devs to once again make the Master fire the T21 faster than a Master Pistoleer can fire an FWG5 Pistol. Ask for some major improvements that help the class as a whole, not just the people who have one box. 2.5 melee reduction, major HAM cost reduction, get rid of the T21 nerf, etc.
Philosopher1976 wrote:Actually you get no speed reduction at all unless you are a Master Rifleman. Let's take a look at the numbers.
For instance,let's sayyou use a 115-333 7.5 T21 rifle, which a Rifleman posted as a "typical" post-nerf T21 in a post recently. As a non-Master Rifleman (4-4-4-4, everything but the Master box) you fire this T21 at the rate of once per5.4 seconds using the now familiar speed equation:
(WeaponSpeed * SpecialDelayMod) - (WeaponSpeed * SpecialDelayMod) * (ProfessionSpeed/100)
And if result < 1.0s , result = 1.0 (speed cap)Because you're not firing faster than once per three seconds anyway, the "nerf" doesn't affect the 4-4-4-4 Rifleman at all. It _does_ effect the Master Rifleman, however. Currently, the Master is firing this rifle once every 1.8 seconds! Right now most of the benefits of Rifleman are in that one box.
Just realize what's going on here. The result is a zero reduction for non-Masters and a 40% DPS reduction for Masters. I think the solution is not to ask to kite other classes in exchange or to ask for the Devs to once again make the Master fire the T21 faster than a Master Pistoleer can fire an FWG5 Pistol. Ask for some major improvements that help the class as a whole, not just the people who have one box. 2.5 melee reduction, major HAM cost reduction, get rid of the T21 nerf, etc.
I'm not sure where you're getting your figures, but you're firing much faster than 5.4s with a 7.5s rifle as a 4-4-4-4 rifleman. At this point your delay is 2.25s using autofire (+70 speed).
Yazule wrote:one last note before I wander away from this thread. Those logs dont mean a thing, they show the gross damage, but no damage for armor. FWG5 is AR0 so you can devide all those damages by half, and then apply the reduction due to resistances on armor. The log only shows gross damage, and anyone who knows about weapons knows that armor makes a big big difference.
The study needs to be much more comprehensive to be worth talking about, otherwise you will just get people massaging numbers to suit their needs and beliefs.
in the end all I have to say is dont cry nerf or you will get it cried back at you. call for a max range nerf on pistols and you will get a min range nerf called on you.
Er, it was stated many times over that this test was only checking accuracy. Accuracy is not damage.