Pitfight!!! - Testing units in Battle Editor

[ddrake]

Hi all,
have been doing some testing with the battle editor (for 1.011 now) by Iberon and Kolba (Thank you! ). Mainly putting in battle armies all made up by a single unit and having them fight each other without any meddling by me.
Testing was done so far on a 80 vs 80 battle. Probably higher numbers would reduce randomness further.
Very large map for ranged units & cavalry, medium map (mountains) for infantry (doesn't change much since they actually fight on plain but they get less time in clashing together).

More or less results replicate what everyone knows.
Some other results are quite surprising (at least to me).

For instance how good Vaegir Guards are. I hardly ever have them in game because they take quite a lot to train. But in my tests they soundly defeated the Rhodok Sergeants (80-43), Swadian Sergeants (80 - 25 !!) and lost at (37 - 80) against Nord Huscarls.

S Sergeants had mixed results against R Sergeants depending much on how much tight the fight was. In close range the Swadians got an advantage.

Nords appear to have the best mid-level infantry too. Also thanks to their big shields and short axes so useful in close combat.

Rhodok Sharpshooters clearly emerged as top ranged unit:
against S. Sharpshooters (80 - 25 !), V. Marksmen (80 - 49) and K. Vet Horse Archer (80 - 19 !!!).
By the way. R Sharpshooters even defeated (with heavy losses) the N Huscarls after I gave them the simple command to hold ground & spread out: 80 - 70.
It must be said though that since the AI just charges ranged units come to melee fight way more than in battles with AI lords.
Considering how easily they can get trained and their cost I think they can be considered THE top unit (I know, now every non-Rhodok will be mad at me  ).

Will do some more testing next, especially if i find a way to have the armies keep formations instead of behaving like a mob 

P.S. I DIDN'T do much testing with the S Knights because everyone knows how good they are.. excluding hilly terrain and sieges... 
Still did a try of the classic S Knights vs N Huscarls in open ground... 80 - 22 for Swadia. There's not much else to add, is there? 

 

[captain lust]

this isn;t really a fair test.

Some troops do well because they are in large numbers (rhoddoks).

Also it's on flat land.

 

[kosmik]

and if your giving orders, that means you are interfering.

 

[Danath]

Ranged units tests are worthless since they always charge. About the loses of the Huscarls vs R. Sharpshooters, it was on plain terrain or mountain?
I once made tests for TEATRC mod and infantery with shields defeated easily ranged units. Since the AI use of shield is kind of random, probably more tests will give many different results

 

[ddrake]

this isn;t really a fair test.

Some troops do well because they are in large numbers (rhoddoks).

Also it's on flat land.

True. 2 R Sharpshooters vs 2 N Huscarls would have no chance. But 80 vs 80 is a reasonable sized battle. Quite common for us lower-tier pc owners 

Actually I will see if I can make a true 'flatland' map. Sloped terrain induces some randomness.

@kosmik: I didn't issue orders. Only in the case of R Sharpshooters vs N Huscarls. In that case it was 2 orders. Just to see how they would perform. And yes, that is not fair. Huscarls came up forming a kind of colomn formation, taking shots from the sides.

@Danath: depends what you want to test on a ranged unit. Knowing it's melee performance is good too.
Ideally one should be able to place the two armies of ranged units facing each other and then give order to start shooting and see how it goes. But that can't be done. Maybe with Warband?
Ranged units battles and cavalry battles will always be on the very large map which is (mostly) flat. Has a central low hill though.
Yep, more fights would reduce random results. It would help if one could write scripts instead of setting each battle in battle editor menus.... (hint mode on..  )

By the way: if you have suggestions on how test should be conducted you are welcome.

Next match: K Vet Horse Archers vs K Lancers. Large map. Will try battles with different battle sizes.

K Lancers kicked K Vet Archers hard winning 3 matches in a row at a staggering 80-8, 80-11, 80-15 score.
Will the Vets prove good at anything? Or will they be declared worst top tier ever?
Next on Sick, Sad W... uhm, no, sorry. Next, on M & B Pitfight channel!

 

[Gabeed]

I think, no matter what, the test will be biased in someway, since the variables of battles vary greatly.  One could argue that not giving orders to troops is just as unfair as giving orders to troops.  Each faction and troop have their niche application, and in my opinion all the "tests" that have been conducted by this OP as well as others are not as factual and clear cut as one might think.

They're still fun to read and observe, though. 

 

[kosmik]

Also the huscarls have ranged weapons and they start throwing them from far away so you would have to tee them to hold fire. But it's great that someone had enough tome to do one of those tests again.

 

[fisheye]

Instead of having them fight each other, which tells you a lot of nothing (how often does the AI have 80 nord huscarls?), you should have the test troops on your side, and fight a few fixed "reference" forces, e.g. all high level infantry, mixed infantry and cavalry, mixed infantry and archers. That way you get a sense of the troop's general versatility, and whether it has particular strengths against certain types of enemies.

 

[duracell]

I think vaegir guards are better in open battle because their armor is not that cumbersome, they have points on athletics and lots of them has two handers but in sieges the swadians are better because of their superior armor.

Lancers defeat horse archers because lancers are the best anti-horse archer unit, this doesnt make the horse archers worthless.

I would like to see the test result of the AI faction armies, like a Rhodok Lord vs. Khergit Lord battle but without any player interference, in autoresolve only numbers and troop levels count and the faction weaknesses and strengths are not represented at all

 

[Zilberfrid]

You know, you can do that by just waiting about in the map...

 

[captain lust]

I think vaegir guards are better in open battle because their armor is not that cumbersome, they have points on athletics and lots of them has two handers but in sieges the swadians are better because of their superior armor.

Lancers defeat horse archers because lancers are the best anti-horse archer unit, this doesnt make the horse archers worthless.

I would like to see the test result of the AI faction armies, like a Rhodok Lord vs. Khergit Lord battle but without any player interference, in autoresolve only numbers and troop levels count and the faction weaknesses and strengths are not represented at all

have you tried the custom battle editor mod?

 

[ddrake]

Hi guys,
I tried some more matches.
First of all some more testing on the K Vets. Moved by Duracell post I tried giving them another chance... boy! they DO suck! Here are the scores:
K H Vets vs R Sharpshooters: 19-80
K H Vets vs R Vet Spearmen : 45-80 (average on 3 matches) oh well, Rhodoks are K eaters... 
K H Vets vs S Sharpshooters : 35-80 dang!
K H Vets vs V Marksmen: 39-80 ! and Vaegirs have no shields! 
K H Vets vs V Knights: 5 - 80 !!!! 
I even tried dismounting them and let them charge V Marksmen. Results were not much different than when mounted: 46-80 
Basically we can say they are a pathetic top-tier unit, can we not? 
But am glad of that. I hate Khergits anyway and having highly effective horse archers would make it extra difficult to beat them...

Some more tests on the Vaegir Guards:
N Warriors vs V Guards: 36-80 ok warriors are tier 4 only...
N Veterans vs V Guards: 53-80 3 matches average
Considering the previous results V Guards show to be pretty strong.. Probably come second in infantry ladder

And finally an all time classic: R Sergeants vs S Sergeants. This time to even out randomness I set armies of 250 units each. Two matches so far.
R Sergeants vs S Sergeants: 250-234
R Sergeants vs S Sergeants: 210-250
Which shows that they are quite balanced, maybe a bit of advantage on the swadians because things tend to get crowded and at close quarters their morningstars are more effective than the glaives.

Tests are with two unit kinds only because using more would add randomness to the results. For instance if there is some cavalry results may be changed by how many enemy infantrimen start running after them leaving their comrades fighting against the bulk of the enemy force. Still maybe i'll do some test with mixed forces.

 

[captain lust]

These tests aren't really valid.

Khergits just require a little more micromanagement than other troops.

Just saying charge is an unfair test.

however, your tests with rhoddok and swadian seargants are quite interesting.

 

[duracell]

Tests are with two unit kinds only because using more would add randomness to the results. For instance if there is some cavalry results may be changed by how many enemy infantrimen start running after them leaving their comrades fighting against the bulk of the enemy force. Still maybe i'll do some test with mixed forces.

And this is exactly why you shouldnt say horse archers suck just because they alone are not that great, they are a key element in khergit armies, horse archers are the ones who skirmish the enemy so the enemy doesnt keep formation, they start to chase the horse archers all over the map -> lancers are not knights, they have crap horses and weak armor/melee skills so they suck in tight melee BUT they are great against disorganized troops so the two troop types helps each other out.

 

[killer-blead]

"K H vets" ? you can just type the full name of the troops your testing that way there won't be any mis understanding.

 

[Ruthven]

"K H vets" ? you can just type the full name of the troops your testing that way there won't be any mis understanding.

No, Khergit Veteran Horse Archers is no fun to type lots because the name is fairly long.

 

[captain lust]

We all undestood straight away killer.

 

[killer-blead]

well i didn't 

the test results weren't that bad for the swadian sergeants. thats pretty weird.

 

[ddrake]

@Cpt Lust: AI does no 'micromanagement'...

@Duracell: this may be a good point. Some troops are there as 'flankers'. But since the AI doesn't manage the complex mechanics of commanding an army of different troops what would I be testing? The outcome would be mostly random.
Re-enacting the typical battles of Calradia is not the point here. You can do it in the game better. There troops are kept in line. Archers do stay back (mostly) and use ranged weapons.
Tests in battle editor seem better for infantry and cavalry units. Not so much for ranged ones.
Still.. losing 5 to 80 is such a low result that Khergits should get out of Calradia in shame! 

@killer-blead: do you really misunderstand it? 
K H Vets = Khergit Veteran Horse Archers
Am just trying to save bandwidth.. you know protect the environment and all that stuff... 

 

[captain lust]

Er drake i realise that.

Its my whole ****ing point you intolerable mong. Under the correct micromanagement of a human, they are a lot more effective than when they are just being told to charge.