LRM testing? 20v15v10/arti vs. not etc

Thread in 'Betty's Playground' started by That Dawg, May 1, 2015.

  1. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    for giggles, I have 2 lrm 15's and 2 lrm 20's in my warhawk

    launched to tourmaline, dont walk, turn right, an atlas- you have the atlas's left torso/arm facing you, the range is the same, torso twist the same, heat the same, only one volley, return to mech lab, repeat. fresh target mech each time.
    tried to keep it as scientifically the same as possible, not "feels like this is that"

    lrm 20-98% lit up half the target with hits
    lrm 15-98.3% about same hits on target
    lrm 10- 99%, one or two less hits per volley. probably great on builds you want assists or slots, WHK is slot poor.

    A hair better, and I mean barely better results with the 20 vs. the 15

    double up- fire TWO at the same time:
    2x20 97% about the same hits as using single lrm
    2x15, 97.25% same hits on target

    based on that...as has been suggested but had to see for myself, the extra weight, heat, ammo consumption doesn't really help buy using LRM20's (clan..maybe IS is different)

    now..this is where I find it very interesting.
    Introduce Artemis- visually, it really does LOOK like more missiles are hitting, the lrms went from chewing up dirt to all hitting.........except the damage was nearly the same.
    The paper dolled showed the same areas hit, the percentages were slightly better(firing x2), lrm20's and arte got 96.5%, lrm15's and arte got 97%
    If I sat there till an arm blew off or I killed the mech, I'm talking arte did the job in one salvo faster..MAYBE, a couple of tests it was so close as to be a judgement call, where arte just felt like it did a better job- cause all the missiles looked like they were hitting.

    I have arte on all I own..always have, cause, its better. and stuff..been told that forever.
    but...the extra weight, slot consumption..making me wonder..the penalty vs. the profit.
    anyone else do such moronic time suck of lrms and care to share? anyone find links supporting the validity for arte or not depending on tubes etc?
    p.s. I am NOT above exploiting mechanics of any game. Hence, the query.
     
  2. enileph

    enileph Star Lord

    4,203
    518
    201
    I suggest instead of doing one volley, you bring ample ammo and count how many volleys it takes to kill the target. That would be a more useful information on damage, rather than the area it hits.

    As for +A
    1) It works great with LRM20 and LRM15. LRM10 is a toss up really, it does improve but a normal LRM10 is still acceptable, especially if there are spread quirks on it.
    2) It is a waste of tonnage and slot for LRM5 which is quite accurate to start with, esspecially on multiple LRM5 builds.
    3) The +A effect is much lower if the target is not directly spotted.
     
    Last edited: May 1, 2015
    dog_funtom likes this.
  3. epikt

    epikt Benefactor

    2,511
    502
    104
    I have a problem with your method: you only shoot one volley, while on the contrary you'd want to shoot the more you can in order to reduce the uncertainty.

    I explain. Since the damage percentage only decrease one point at a time, the uncertainty is one point (Captain Obvious!). The damage is 3 points at max, that means the uncertainty is 33%! That's HUGE.
    If you shoot 10 volleys the damage would be approx 30 points and the uncertainty still 1 point, making it 3%. Much better.

    edit: it would also make reading the paper doll much easier.

    Also, here's a good read: http://www.mechspecs.com/threads/system-test-lrms.8298/
    (only tested on IS iirc)
     
    enileph likes this.
  4. epikt

    epikt Benefactor

    2,511
    502
    104
    Artemis has no effect at all on indirect targets.
    (according to the official "tips" and any tests I read)

    edit: but yep, as a basic rule no artemis on 5s and 10s, artemis on 15s and 20s. Though I don't know much about clan missiles.
     
  5. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    I did several volleys, with each "test", then took the average of upwards of 5 "attacks". I have done the how many volleys it takes to kill a mech and frankly disillusioned from the results i.e. 7 to 8 volleys to kill regardless, and I was more interested in damage on undamaged mechs for this test

    as said, I ran it several times, took the average, this was not fire once, record and take as fact.


    most of my lrm targets are indirect, but I could see the atlas in my tests.
     
  6. epikt

    epikt Benefactor

    2,511
    502
    104
    I'm afraid I don't understand. In your original post you wrote:
     
  7. tfun90

    tfun90 Advanced Member

    498
    110
    43
    dog_funtom likes this.
  8. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    ah @ epikt In order to get the results-I would launch one volley at the mech- record the damage, reset the map/mech and do it again- wanting to see damage on fresh mech as the standard.

    too many times in the past, I would sit there and launch 2-3 volleys to see what damage.
    If you waste your time (testing is a pain) I started to see differing results with each new shot fired.
    lasers/ballistics/lrms take five shots at something and I was rarely getting repeatable results.
    1st shot X damage, 2nd 3rd 4th shot, would often be different.
    2nd shot sometimes would knock the mech down 3%, sometime 4% sometimes 1%
    Taking the first shot was always the same, see what I'm trying to say?

    What mechanics the game uses for subsequent damage, I could not get repeatable results. Maybe I walked, or moved my torso, but something changed.

    so, hoping to eliminate that- spawn, first shot, record, respawn repeat 3-5 times for an average.
     
  9. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    hell, wish the search showed that when I looked, thanks
     
  10. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    nice link, covered quite a bit of what I was theorizing about, especially with arti vs. not.

    well this is maddening

    did the volley test....
    warhawk
    didin't matter, 2xlrm20, 2xlrm15 OR 4xlrm15 8-9 volleys to kill the atlas
    NO difference. the 2xlrm20 knocked its arm off pretty steady on shot #4, but not always volley #5 would knock off the arm with either build
    ....4x lrm10? 10 to kill it volley 5 or 6 knocked off the arm
    all tests without arti...and I'm going broke buying, removing, buying removing the arty derpaherp

    that other guys testing helped explain why my kintro 18 with lrm5's was so deadly, one of the few IS mechs I kept

    so...maybe the warhawk with 10's might do as well in battle as the 4xlrm15 setup and allow more heatsinks
     
    Last edited: May 1, 2015
  11. epikt

    epikt Benefactor

    2,511
    502
    104
    Your target damage percentage after the first salvo being consistently the same does not mean the damage dealt has been consistently the same. It only means your measurement tool is not precise enough to measure the variations.
    You can't improve your tool's precision, but you can "distribute" this imprecision over multiple measures, by shooting multiple times and calculate averages (*).
    There is a random factor in LRM hit distribution so there will have variations, you can't dismiss test results because they show variations. But again, the more salvos you're measuring, the more you can apprehend those variations.
    Hence what I wrote in my first post: as many salvos as possible are required to get reliable results.

    (*) this only works because what you're measuring is stacking shot after shot while the imprecision is constant

    I went to the testing ground to test myself, and my first thought was your position is pretty bad. You don't see your target completely (actually, my first drop was on an Adder and I didn't see it at all from the spawn) and more important you're shooting at it from the side (which shouldn't be such an issue for the first volleys, but as soon as the arm is destroyed and thus the hitbox changed you're screwed, not to mention that when the side torso is destroyed all damage dealt to this part is reduced). You'd want a position facing your target. In addition to allow you more volleys on the same hitbox, it will also allow you to study the spread of the missiles (like Skribs did).
    That being said, the damages I got looked consistent enough (5 sessions with 2 LRM20s). But I won't continue with those settings - like you said testing is a pain, I'd rather do it right.
     
    dog_funtom likes this.
  12. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    That I do agree with. Testing how many salvos to knock down a mech would be but one way to test.

    My goal was to test primary damage, first impact, the difference between lrm 5, lrm 10, lrm 15, lrm 20, and the value as it relates to heat, weight and slot usage.
    Immediately its clear (as other have said) lrm 15's are better than lrm 20's, not so much because the damage is the same, but its SO close while using less slots, less weight, and consumes less ammo with each single volley.

    uhm...in my warhawk, I see 3/4's of the target, and I dont move- you want to make your self nuts? MOVE and see see different results.
    When you move you just changed the one true constant. Just changed your elevation, distance, angle/position to target. I'd suggest finding a map that midget mech can actually see something. I think youre too short for caustic, I'll check some later on and find one.

    I rarely got to face the enemy when launching lrms, mostly I wind up being on the side, or they torso twist.

    yup, same goes if you're standing in front of it, or behind it, and knock a leg off, or an arm. hit reg is some bizzare thing, and you can't control it.
    See how many volleys it takes to kill the atlas from that side with lrm15's, then use lrm20's, then go to the front, you'll go insane trying to clinically figure it out.
    I was getting 8-10 volleys to kill it-regardless BUT that is precisely what I wanted to know!...10's took more 20's took less, but the 10's were consistent, 20's did it 7 or 8, 15's did it in 8...all but once..

    anyhow, the warhawk spawned my need to know, limited slots, limited weight, squishy sides, lrm quirks, either it becomes my clan lrm boat aka the old stalkers, or I sell it.

    and what I found out? in summary, the lrm10's almost do as good a job as the lrm 20's, ammo lasts longer, the damage is soooo flipping close, the extra slots and weight I PICK UP allow me bigger energy weapons.
    Big picture? warhawk with 4xlrm10's and more heat sinks, 5MPL's or 2ERLL and heatsinks is more proficient than warhawk with 2xlrm 15, 2xlrm20 and four mediums.
    your mileage may vary, and all results may change next patch......ffs....
     
  13. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    oh, btw, I run my adder in pugs with lrm15 3ML's, but curiosity now, I might go lrm10's. I will try and find a map that spawns better for short mechs and use that for all testing.

    I would LOVE to see the actual algorithm they use in hitreg and if its the same across the board, or changes from weight class of mech, or IS to Clan

    you knew that Artemis install on streaks is an exploit?
    No weight penalty but you get the bonus for faster data feed, lock times etc.
     
  14. epikt

    epikt Benefactor

    2,511
    502
    104
    So you did shoot series of volleys afterall...
    You don't have to worry with most of my rant then.

    Wrong.
    First when you're facing the target most of the missiles go to the CT, which is destroyed first (cf Skribs' post).
    Second the hitreg is much better (if not perfect) on the training ground since it's all client-side. In fact even on an actual server when you shoot missiles at a facing target (dumb enough not to torso twist) you will destroy the CT first.

    Btw you're not even interested in taking down that mech, you're much interested in the damage dealt and damage spread on the hits before your target is destroyed. Usually the final volley screw your measurements, because it might have took one missile just like 20 (that why a one volley difference means nothing)(in tests). What's interesting about the Atlas is the fact it's the closer you have from a good old shooting target. With this you can appreciate damage dealt and damage spread, because the target is symmetrical and constant over time)

    But if you're satisfied with the answers you get even if your test protocol is flawed, what can I say?
     
  15. That Dawg

    That Dawg Well-Known Member

    185
    40
    37
    I dont worry about anyone's rant. I do not mind, ever, being wrong. I'm more interested in my game play, and appreciate someone taking the time to fuck around with it.

    We've had a few patches since January, one directed at hit reg.*

    then...why not remove that variable from your testing?

    I was not, am not, wont be concerned with how many volleys it took to take down a mech, but strictly comparing 5's, 10's, 15's, 20's how much damage the FIRST volley did, based on a series of averages taken from known, repeatable positions at a known fixed target.
    Successfully removing the games variables by walking around, re-positioning etc.

    Then, taking that data and working up a build for a new mech that has limited slots to work with.
    I am utterly satisfied with my results, and perhaps something in this thread will be useful to others.

    still, a good debate is rewarding, so back to the lab

    *that still seems accurate, attack from the side, and moving to exactly the same distance, in front took aprox half the salvo's.
    That could very well make the 20's better than the 15's, better than the 10's

    IF you can get the target to face you directly, not move, it will take less hits. assuming no torso movement etc etc

    only did one run, weekends busy.
     
top-fast
top-fast
top-fast
top-fast