Anyone interested in doing some tests?

[Sasaki Kojiro]

I've read several posts here that state that Weapon upgrades do not affect the damage projectiles do, they only affect h2h. This is what I always thought and it makes more sense to me. However, I have seen quite a few people give their ranged units weapon upgrades and they claim to have better results this way. I tried to do some tests vs the computer but it failed to cooperate, it would use it's archers in h2h and rout it's troops off the map, or march them in at an angle.

I would like a definitive answer to this and so if anyone who can host a game would post a time to meet up with me in the foyer for some quick tests I'd be much obliged.

And if anyone who has done tests like this vs the computer would care to tell how they did it that would be helpful as well.

[TenkiSoratoti]

Well it depends, doesn't it? Are you trying to work this out for SinglePlayer or Multiplayer? The two variables here have alot of differences. I suspect you want this to be solved with Multiplayer, so i'm afraid SinglePlayer is kinda out of the equation. I never really use weapon upgrades for my projectiles, they can become very expensive and I know as much as to say you need to have two weapon upgrades to make any kind of difference.
It's a pitty because most nights I see you in the Lobby by yourself but niether of us can host. I can host for MTW fine and that's more senstitive to router's than STW is

Oh well, life's unfair.

TS

[ShinGaiJin]

I volunteer for your testing Shinano,Ive actually done a bit of research myself on this..but lets get together and try some stuff out

ShinGaiJin

[Puzz3D]

I ran ranged unit tests with two machines on a LAN last year. I'll see if I can dig up the results and post them.

[Sasaki Kojiro]

This is your test results...

"I just ran a test on this. 60 man units, on green, fine weather, firing on YS in close formation just inside max range. The SA fired 28 arrows/man. The musk fired 1 round/man.

SA H0 W0 48 kills
SA H2 W0 52 kills
SA H2 W3 47 kills
SA H9 W0 52 kills

Musk H0 W0 15 kills
Musk H2 W0 15 kills
Musk H2 W3 12 kills
Musk H9 W0 18 kills

I'd say upgrades don't help ranged fire. Upgrades do add one melee point for each upgrade. Incidentally, multiplying 15 kills by 40 rounds gives 600. That's how many men a musk can kill if the enemy just stands there in close formation."

MizuYuuki ~~~
Clan Takiyama ~~~

[Sasaki Kojiro]

I just did a test with Shingaijin with h2 W0 sa under same conditions and got 44 kills. You had 52 kills...which is a difference of 8 kills... http://www.totalwar.org/forum/non-cgi/emoticons/confused.gif I guess it takes more testing to get an accurate result.

[Puzz3D]

Shinano,

HeHe. I was kind of shocked for a second there because I was holding the paper with those results in my hand as I read your post which I thought were your results.

There are problems with that test, and it should be repeated in MP. That test was done in custom battle, and there is some indication that custom battle and MP/LAN give different results. I got the YS to stand still by setting the unit speed to 0 in troopstats.txt. Also, the YS were in roughly 8x8 formation, and that's a problem when you kill nearly all the men because the depth of the formation becomes very thin. The kills drop off on thin formations because the arrows over/under shoot. That effect will compress whatever difference was there at the start. With complete control of the target you could put it in 30x2 deep to minimize the varying kill/depth effect. You could also use the higher armored naginata. I also don't see any indication that I made several runs and averaged them to reduce statistical error. I was probably just looking for a large effect at the time. With the guns, I reduced ammo to 1 so I didn't kill all the men in the target.

I think it's well worth repeating the test online. At the time, there were vets who insisted that raising honor helped archers get more kills, and one argument ran that it increased up to h3 and then flattened out above that.

There is a map called ranges-1 that Kraellin made which allows mesurement of shooting range by using tiles of different color. That allows you to set up multiple units side by side at the exact same range and also to repeat tests with the same distances at another time.

I also found a custom battle test of SA vs YA of varying armor for WE/MI v1.02. The SA are H2, 60 men, 28 ammo vs YA of 120 men, 20x6 deep. This is fine day, SA firing at max range (5000/2.5 tiles) in 3 ranks.

Armor/Kills:

0/111
1/98
2/79
3/68
4/56
5/50
6/36
7/33

Again, the fact that it doesn't plot as a staight line might be due to lower killing efficiency as the unit becomes thinner. I did verify lower killing efficiency on thin units in some very carefully run MTW tests with lots of runs with a mesured statistical error less than the effect being observed.

[Mori Zennousha]

greetings, Mimesaka...

since you pointed out to me that you were pretty sure that Weapons upgrades only affected h2h combat, i have looked into it a bit more myself... as far as the "official" game docs go, i can't find anything that suggests that Weapons upgrades increase the kill ratios for ranged units, except in h2h combat, as you said... now, i can't find anything that says it doesn't either, but that may just be a given...

as far as the testing goes, i think more should be done to be certain... further, i think that, while perfectly flat, clear conditions make sense for a "baseline" test, it would be interesting to determine if an H5 SA takes more advantage of height than say an H2 SA, and if so, to what degree... or if an H5 W2 SA is less disadvantaged by rain than an H0 W0 SA is, again to what degree... there could be little "secrets" hidden in these type of tests that could mean a great deal on the Battlefield.... (just a theoretical example: suppose that H6 SA's only get ~1-2 % more kills than H0 SA's on Ironing Board, but the same SA's get say ~20 % more kills than their H0 counterparts when firng down a steep hill... the implications could be far-reaching... )

in any event, i think that many types of tests should be run, with at least 10 repeats of each type of test for statistical stability, or at least some degree of it, as Puzz3D pointed out... i'd be happy to help you do some of these tests... PM me and we'll try to set up some times...

Sayonara,
Z
*bows*

[Sasaki Kojiro]

Puzz3d: The yari sam in my test were in 4x15 formation and I did notice a decrease in kills once the formation got thinner, perhaps this is the reason for the difference in kills, it makes sense that a deeper formation would have more losses.

Zenn: I'll be on at my normal times everyday, so come online then for tests.

[Sasaki Kojiro]

I did the tests in custom battle today. Ys in 4X15 formation, Sam archers in 3X20. Each SA fired 10 volleys. Ironing board, fine day, 1.03, maximum ranged. I tried to make sure the were always head on but they were sometimes at a slight angle. Did three tests and averaged results.

Hon0: killed 22
Hon0W3: killed 22
Hon2: killed 23
Hon9: killed 23

So it appears weapon upgrades have no affect and Honor upgrades have a small effect (if they had fired all 36 arrows then would have killed 3 or 4 more).