H2 WM vs H3 +1W ND... who is better

[tootee]

These two are nearly identical in values (ND 10koku cheaper). So, which do you think is better for multiplayers?

------------------
tootee the toothless warrior (aka goldfish shimazu)
------------------

[Major Robert Dump]

Tootee it really depends on the scenario and the angle at which the units engage each other, as well as what the rest of both armies are up and their positioning. I would definitely say the Dachi would win if they could hit the monks flank or rear, but u also have to account for morale effects of the rest of the fighting

[tootee]

I think hitting an enemy unit's flank or rear with any of these two units will deal a deadly blown. I am more interested in cases where these two units confront each other frontally. I wonder which has a higher chance of winning. If they are about equal, what is the point of buying WM when I can have a higher morale unit at 10 koku cheaper...

--tootee aka goldfish shimazu

[dagdriver]

It's easy:

Try it out in custom battles!

[barocca]

on flatdemo,
march towards enemy from beginning to equalise fatigue.
charge as soon as opponent charges,

and it's monk 10/10 for AI monks
and monks 10/10 for human monks

...

------------------
DoragonBarocca of Clan Doragon

[tootee]

It is not realistic in custom battle because the difficulty level modify the units' strength beyond the default troopstats. So it will not give a fair judgement... do I set it to easy, hard or difficult? I have tried out custom battles on different settings and each give me different result.

Maybe someone with much online experience can tell me...

--tootee aka goldfish shimazu

[tootee]

Quote Originally posted by barocca:
on flatdemo,
march towards enemy from beginning to equalise fatigue.
charge as soon as opponent charges,

and it's monk 10/10 for AI monks
and monks 10/10 for human monks

...
[/QUOTE]

so you have done that?! which is better?

--tootee aka goldfish shimazu

[Vanya]

Quote Originally posted by tootee:
so you have done that?! which is better?

--tootee aka goldfish shimazu[/QUOTE]

My guess is that the whole 'holier than thou' attitude of monks in general tips the balance in their favor over their big-bladed counterparts...

But only for that reason.

No-Dachi have same morale as Monks Vanya.
And same armour too.

Therefore:

NoDachi HON3 have better morale than monks HON2.
NoDachi HON3 have same armour as Monks HON2.

Now...there are many other factors...
I'll check on Multiplayer with a m8...best of 5.

Tera

------------------
Honour to Clan Torîi Aku.

Visit my resource centre here!
Evil is within us...

[This message has been edited by Terazawa Tokugawa (edited 10-09-2001).]

Ok tried and tested.

Out of 5 battles, the NODS honour3 wep 1 WON 4 of them. The nods have a really good charge bonus now...YES they are better than monks honour2.

Tera

------------------
Honour to Clan Torîi Aku.

Visit my resource centre here!
Evil is within us...

[Major Robert Dump]

Tootee, in the scale of a larger war the unit types must be accounted for.

No Dachi are armorless shock troops. They die big but also take a lot of guys with them when they go. Think back to your really big battles...how common is it to see a coiuple of Dachi units running around with about 10 guys remaining? I see this all the time because they simply dont last in prolonged battles (unless they are kept in reserves the whole time) Whereas monks will often go into a battle and come out half stong or more

I would say monks have equal figthing ability, but are harder to knock off due to the morale effect and their higher armor.

My policy is to use several Dachi and one or two monks on attacks, but never use them on defense unless i expect to be rushing my attacker. Monks are the better choice defensively.

Monks can fare better defensively vs other units.

But when confronted head to head, a NOD honour3 with WEP1 will quite always win the monk, as I said before.

Tera

------------------
Honour to Clan Torîi Aku.

Visit my resource centre here!
Evil is within us...

[Shoko]

What about in the old STW?

[Gothmog]

In old STW, WM will win big time.

The Wep upgrade does most of the job.

In STW you don't have upgrades, and neither an honour4 nodachi would beat a monk probably.

Tera

------------------
Honour to Clan Torîi Aku.

Visit my resource centre here!
Evil is within us...

[Vanya]

I see you have chosen 'The Big Sword' over the 'Good package'...

[tootee]

Thanks so much Tera, Vanya, Dump and guys! I guessed ND H3 +1W is better over WM H2 in multiplayer 1-on-1 (in custom battle under expert mode, they are evenly matched, with me controlling ND. I guess when it translates to multiplayer mode ND is better).

But Tera, you mentioned that ND has same armour as WM? Probably I need to look at the troopstat again.. I only know that ND has -ve modifier for defence? meaning the opponent has a higher chance to hit? Is the armour rating added to the chance to hit by opponent , or is it used to determine whether a hit is fatal? You get what I mean?..

------------------
tootee the toothless warrior (aka goldfish shimazu)
------------------


[This message has been edited by tootee (edited 10-10-2001).]

[Khan7]

In the actual troopstats values, only melee bonus affects chance to hit in melee, only defense bonus affects chance of being hit in melee, and only armor bonus affects chance of being penetrated by a projectile.

In melee, the attack bonus of the attacking sprite is checked vs. the defense bonus of the defending sprite to determine outcome. Every hit is considered a kill.

With missles, a projectiles must first hit, and then check vs. the armor value to kill or go kapwing.

Of course armor upgrades, in the game, affect both defense bonus and armor level. But in troopstats.txt, fortunately, it is handled nicely and cleanly in the way I just described.

Matt

[Gothmog]

Quote Originally posted by Khan7:
In the actual troopstats values, only melee bonus affects chance to hit in melee, only defense bonus affects chance of being hit in melee, and only armor bonus affects chance of being penetrated by a projectile.

In melee, the attack bonus of the attacking sprite is checked vs. the defense bonus of the defending sprite to determine outcome. Every hit is considered a kill.

With missles, a projectiles must first hit, and then check vs. the armor value to kill or go kapwing.

Of course armor upgrades, in the game, affect both defense bonus and armor level. But in troopstats.txt, fortunately, it is handled nicely and cleanly in the way I just described.

Matt[/QUOTE]

That's very informative, Khan. I have always been wondering what's the diffrence between the rating of armor and defense ...

[Puzz3D]

tootee,

Use normal difficulty in custom battle. The ai gets advantages on hard and expert.

khan7,

The chance to kill in melee is attack - defense + bonus. So the striking man's attack value combines with the defending man's defense value to determine the chance. Then it switches around as the man who was defending strikes at the man who just attacked him. I never tried to figure out the absolute probability that zero represents, but it could be done experimentally. Each point of deviation from zero represents a 20% change in probability according to one of the answers given in EradoSan's Question Time with CA/DT, IIRC.

tootee,

An H2 WM is 6/3, and an H3 W1 ND is 9/-1. This would say the WM has a one point advantage (20%) over the ND as long as there are not combat bonuses in effect. I haven't actually tested these units in WE/MI v1.01 so I'm not questioning Terazawa's experience which suggests otherwise. With the hidden bonuses being given to some units, you have to go with experience rather than calculation. You should run the experiment many times, preferably with 120 man units to minimize statistical fluctuation, on the green map with both units charging, both on engage at will and both at the same fatigue level. It's best to test online, because in custom sometimes the ai will do things with a unit that ruins the test.

MizuYuuki ~~~
Clan TAkiyama ~~~

I tested online, fine day, green, with units marching at the same time...and same fatigue level. On 4 occasions out of 5 the Nodachi won...and the one lost was due to the death of the Nodachi general immediately...

Probably there is more...but something is for sure...it's was always a close battle.

Monks aint unbeatable...

Tera

------------------
Honour to Clan Torîi Aku.

Visit my resource centre here!
Evil is within us...

[Khan7]

Hmmm.. I have recently disliked 120-man units..

Because you see, due to the sheer dizzying quantity of testing I must do, and the schedule I must do it on (in perusing possible modification to the game), doing it with a buddy isn't feasible. So I am stuck with the imperfect system of SP Custom battles.

One very important thing to note-- the AI *DOES NOT GET TIRED*. They never suffer from even the slightest, smigdiest bit of fatigue. This necessitates testing from both sides of an equation, and also, hypothetically, could lead to another phenomenon I've noticed--

120-man units in my experience wreck tests. The dynamic is simply different. What I hypothesize based on my observations is that 120-man combat takes longer to resolve, so therefore the fatigue factor cathces up with the non-AI unit to a greater degree than it does with 60-man units.

But there is a profound difference, you must believe me, at least in 1v1 tests. It's as big a difference as a unit getting beat 9/10 with 120-man units, and winning 9/10 with 60-man units.

And I can't always test from both sides, simply because of the uncooperative behavior of the AI.

.

It is also important to note that it will be very very difficult to experimentally determine percentage values. Simply because if a unit has a SLIGHT advantage, you will expect it to win 4/5 times on average. I have my doubts as to whether the effects of a single point difference could be determined reliably, except under the most immaculate lab conditions and with the most tedious and numerous schedule of repeated tests.

Matt

[tootee]

Quote Originally posted by Puzz3D:

The chance to kill in melee is attack - defense + bonus. So the striking man's attack value combines with the defending man's defense value to determine the chance. Then it switches around as the man who was defending strikes at the man who just attacked him. Each point of deviation from zero represents a 20% change in probability according to one of the answers given in EradoSan's Question Time with CA/DT, IIRC.

An H2 WM is 6/3, and an H3 W1 ND is 9/-1. This would say the WM has a one point advantage (20%) over the ND as long as there are not combat bonuses in effect.

MizuYuuki ~~~
Clan TAkiyama ~~~[/QUOTE]

Geesh, Thx Yuuki for the great info. BTW where can I find these info online (things like how you derive 6/3 for WM H2)? You mentioned Erado San has a Q&A with CA/DT?

I think to be good in MI online is to play alot online, and to learn every lesson from defeats, own and enemies.. but knowing the mechanics of the game probably gives a very small edge over the opponents who do not, when things are equally matched

I will redo my experiment in custom under normal difficulty. Probably it doesn't mean much at the end of the day :P

------------------
tootee the toothless warrior (aka goldfish shimazu)
------------------

[Puzz3D]

tootee,

You can look at an online Strategy Guide for STW here: http://gamespot.com/gamespot/guides/pc/shogun/

You have to look in the TroopStats.txt file in your main WE/MI directory to see the parameters for the new game. Some units, like MHC, are getting hidden bonuses that you cannot see.

Khan7,

Your observation about the ai not suffering fatigue would explain the small difference I observe when testing both ways against the ai instead of using a LAN. It would be more pronounced with 120 man units. You could turn off fatigue to eliminate this effect, but different unit types might fatigue at different rates while fighting, and you would be eliminating this factor from the test. The units fatigue at different rates while moving, but I don't know about fighting. I think you are right that results from custom battle testing are suspect at best.

When you run the same matchup many times, keep track of the total kills, and not just the win/loss results. If you look at the ratio of total kills, you reduce the statistical noise more than just looking at wins/losses, and you don't have to run the test as many times.

I always watch the casualties build up in both units in the first part of the battle to get a feeling for how much a unit is better because, once a unit gets a numerical advantage, the extra men start ganging up on their adversaries. So, another factor comes progressively into play where a defending man can be struck at more than once, but he can only strike back at one of the men attacking him in that cycle. The numerically superior unit will get an accelerated number of kills.

MizuYuuki ~~~
Clan Takiyama ~~~

[tootee]

thx so much! this is what I have been looking for.

------------------
tootee the toothless warrior (aka goldfish shimazu)
------------------