Terrain testing results
Printed From: Illyriad
Category: Strategies, Guides & Help
Forum Name: Strategies, Tips & Tricks
Forum Description: Player created guides and advice.
URL: http://forum.illyriad.co.uk/forum_posts.asp?TID=961
Printed Date: 17 Apr 2022 at 12:25
Software Version: Web Wiz Forums 12.03 - http://www.webwizforums.com
Topic: Terrain testing results
Posted By: KillerPoodle
Subject: Terrain testing results
Date Posted: 27 Aug 2010 at 16:02
|
After some debate in Global about the impact of terrain, Zangi and I decided to do some controlled tests. We agreed to use 100 basic dwarf bowmen (99 Slingers plus unbonused commander) vs 20 Human cavalry (19 + unbonused commander). This gives a nice comparison because 20 cavalry have an attack strength of 1300 (65*20) and Slingers have a defence against cavalry of 13 so the armies are equal when the cavalry are attacking the slingers. First test was on a mountain square type "Lonely Peak":
As you can see - rather than this being a close fight the terrain gave a significant advantage to the bows. |
Replies:
Posted By: KillerPoodle
Date Posted: 27 Aug 2010 at 16:19
|
I'd love for everyone to add to this - I suggest you try to keep the armies involved as simple as possible and refer to this for stats to decide numbers: http://uk1.illyriad.co.uk/view_all_units.asp?unit=0 |
Posted By: KillerPoodle
Date Posted: 27 Aug 2010 at 16:43
|
Some math: a = cav atk terrain modifier for mountains b = bow defence terrain mod for mountains My cavalry killed 53 out of 99 troops (and some percentage of the commander) so I'm going to assume our relative strengths were 53/99 = .535 My strength was 1300*a and Zangi's was 1300*b thus: 1300a/1300b = 0.535 and; a/b = 0.535 So, for example that could give values for a and b as follows: 1.0 and 1.869 or 0.8 and 1.49 (e.g. a 20% penalty to cav atk and a 50% bonus to bow defence) or 0.6 and 1.12 etc... There are an infinite number of possible values from this one equation - we would need more results to start narrowing down the actual values. |
Posted By: KillerPoodle
Date Posted: 28 Aug 2010 at 04:50
|
Posted By: sityviper
Date Posted: 28 Aug 2010 at 17:20
|
you have to also figue on who is defending, defence has the bonus ------------- §.i.†.¥.V.‡.¶.€.® 3 |
Posted By: Mandarins31
Date Posted: 28 Aug 2010 at 23:52
|
Hi, you may consider that on the same square and for the same unit, def and attack bonus could be different. for exemple if on lonely peaks bows have +50% def, maybe they dont have +50% attck but a diferent %age of bonus |
Posted By: Shrapnel
Date Posted: 29 Aug 2010 at 01:42
|
Two opposing forces clash against each other. Cavalry comes into its own when able to strike hostile forces at will, and from unexpected directions - and nowhere is this more feasible than on open plains. Lightly armoured spear units, however, prefer terrain where there's some cover available. Fighting defensively on open plains, cavalry draws strength from the ability to form and reform their lines of engagement depending on the direction of battle, and it is here where cavalry excels.
In this experiment, infantry went against spear men. The attack value of the infantry equals the defense value of the spear men. The battle happened on the plains. These results are surprising to me. If we assume there is a random element, then I'm thinking neither side had a bonus or penalty thus resulting in the 7 surviving axe men. Either there is a random element and no modifiers, or no random element and just a minor penalty to the spear men? I really thought the axe men would have a lot more survivors. |
||||||||||||||||||||||||||||||
Posted By: Mandarins31
Date Posted: 29 Aug 2010 at 02:40
| here infantry has a non negligable advantage on spearmen. they have 15-20% of survivors (7/39*100) while all the speamen are dead. so infantry has a medium advantage when attacking spearmen on plains. |
Posted By: KillerPoodle
Date Posted: 29 Aug 2010 at 03:37
|
"
Lightly armoured spear units, however, prefer terrain where there's some
cover available." Maybe spears have a penalty on plains where there is no cover? BTW - I think you can get a more accurate picture when small numbers of troops are involved by looking at the damage to the commander (84% in this case - so 16% remaining). |
Posted By: Mandarins31
Date Posted: 29 Aug 2010 at 03:52
for me the only way to know the %age is first of all to find a square on which one an unit is not affected... so +0% bonus on attack or defense : so we could have the (Power)*1.0 and find the %age value of the other units on this square. i think about plains and swordmen, its easy to know the bonus of every units on it because swordmen are unaffected: Hills also benefit ranged units and spearmen to a lesser extent, and still penalise cavalry, but also to a lesser extent. Swordsmen are largely unaffected by hilly terrain. Plains provide the opportunity for cavalry to shine - their ability to maneuver makes them the masters of open flat space. Lightly armoured spearmen dislike the open terrain of plains, where they are easier targets for cavalry and archers. Edit: swordmen largely unaffected on hill (they are affected on mountains). they dont talk about them on plains. we can strongly suppose that swordmen are totally unaffected on plains It will be much difficult on hills, forests, and mountains because we dont have any reference: every unit seems to be affected by the type of these grounds -Exemple on plains with the previous report (swordmen attack spearmen): they both have a power of 1000 (for exemple) a=%bonus on swordmen attck b=%bonus on spearmen def 39-7= 32 32/39= 0.821 1000*b/1000*a=0.821 b/a=0.821 swordmen unaffected => a=1.0 b=0.821a b=0.821 => 17.9% penalty on spearmen def  (but because of the weight of the army, the luck etc... its better to say a penalty between -15% and -20%) |
Posted By: Akita
Date Posted: 29 Aug 2010 at 05:46
|
Are you sure you are applying the appropriate methodology for testing this ? I have never inquired about the exact mechanics of battle (actual bonuses, damages dealt, hitpoints and numbers of casualties resulting), so I am by no means saying this with any "staff" authority, but merely from the perspective of a regular player (so I might be completely wrong). I have at least on one occasion noticed something that appeared at least a bit strange : a group of high-heroism commanders (no vitality upgrade) with just one regular troop so you can send an army out (raw attack power of a couple hundred commander-type troops) could end up in the death of everything even if the enemy is a wimpy group of low-grade NPC animals, while adding just a few more regular units (which would not noticeably alter the total attack power) will result in just some casualties and army survival. One could conclude that the formula calculating casualties is a bit more complicate than straightforward army attack vs army defense in percentages, and hitpoints might also matter. Then, there's the matter of how many hitpoints do regular troops have (if any, and if it differs from troop to troop). And of course, could always expect to also see some randomness involved, which might throw off reverse-engineering attempts even further. |
Posted By: fluffy
Date Posted: 29 Aug 2010 at 06:19
| wouldn't the easiest way to see about defensive terrain bonuses be to attack and defend with the same type and number of troops? like 100 spears vs 100 spears? I don't know if that would help though |
Posted By: KillerPoodle
Date Posted: 29 Aug 2010 at 06:21
|
There was a change a few months back such that high leveled commanders only had as much attack strength as the sum of the attack strength of the troops in the army regardless of their heroism level. E.g. for a fully leveled knight commander you need 60+ knights in the army to get the commanders full attack strength. |
Posted By: Mandarins31
Date Posted: 29 Aug 2010 at 07:13
|
To fluffy: we cant find the %age of bonus by sending 2 armies of the same type because they will be affected by the same bonus. and much, if the 2 armies have different powers (like in 100 spears vs 100 spears), there is the problem that the bigger the army is, the les casualties it suffers (if an army have a power of 100 and the other a power of 200, to make it simple, the army with 200 power wont lose 100 but less than 100 power). But if we do a battle with the same type of units and the same strenght, we could know if a same unit has the same bonus to its attack and to its defence on this square... |
Posted By: Akita
Date Posted: 29 Aug 2010 at 10:18
|
Â
Ah, nevermind then, that fully explains it. |
Posted By: Shrapnel
Date Posted: 31 Aug 2010 at 01:56
Zangi and I repeated the same experiment to see if there was a random element. We did it exactly the same way and got exactly the same results. This leads me to believe the spearmen did get a minor penalty and that the results are not random.
|
||||||||||||||||||||||||||||||||||||||||||
Posted By: KarL Aegis
Date Posted: 01 Sep 2010 at 01:41
------------- I am not amused. |
Posted By: xilla
Date Posted: 01 Sep 2010 at 09:23
|
As others have stated, I don't think its wise to rule out the attack vs same unit type, but in the same manner as above, calculating att and def values for an even battle. It may very well be the case that troops have the same bonus/penalty in both attack and defence roles, but it's best to rule out all assumptions, especially if you're testing a system that doesn't create that definition. Assuming different att and def bonuses for same troop type, there are 8 variables per tile type, giving 16 ratios that are the only thing calculable from this research. Would still give you the end result, what's better where. How would you narrow down the actual figure though? 16 ratios, 3 battles per ratio per tile for a good test, *currently* 35 different terrain types = 1680 test battles. Good luck :P |
Posted By: Mandarins31
Date Posted: 01 Sep 2010 at 17:31
|
There are only 4 terrain type. Mountains, Hills, Forests and Plains. I don(t think that Lonely peaks, for exemple, gives more Def bonus to Archers than Treacherous Mountains. But yes finally we dont know, and we may be careful of that. |
Posted By: Thexion
Date Posted: 01 Sep 2010 at 20:10
| I think I heard Stormcrow say that there is difference between hills, small hills and so on also with forest and mountains you can read from descriptions that there is already differences. And don't forget city squares under ruined towers ;) |
Posted By: Mandarins31
Date Posted: 01 Sep 2010 at 20:17
|
ok so, as Xilla said that does many tests and i dont know if someone has already found a way to find the bonus of each unit, on each square... actually, for me, the only bonuses we can find are the bonuses of of each units on plains (if we suppose that swordmen receive no bonus on plains). |
Posted By: xilla
Date Posted: 01 Sep 2010 at 21:34
If only we could get a GM to tell us whether or not the modifier is the same in att and def, and if a troop isn't mentioned in the description then do they have a modifier or not. |
Posted By: bartimeus
Date Posted: 01 Sep 2010 at 23:25
|
Part of the fun is to discover that ourselves. ------------- Bartimeus, your very best friend. |
Posted By: Mandarins31
Date Posted: 01 Sep 2010 at 23:25
|
hehe ok so that's a piece of evidence to know if there is a difference between atk and def bonus for an unit, u just need to do a battle on a square with the same power for attackers and defenders of a same type. then this is a so close battle: no modifier difference. if not u can just say if this unit have more def or more atk on this square. if i had a question to ask to a GM that would be this one for exemple: what is the defense modifer of archers on the "lesser" and the "greater" version of each type of square? if only we had those 8 numbers, we could find the modifier on atk and def for each unit on each square. But that may be a too big ask |
Posted By: xilla
Date Posted: 02 Sep 2010 at 03:11
Might be the same on one square, have to do 4 sets of battles with each troop type. Thats 48 battles to prove it for all units on all terrain. Plus, whose to say the t1 and t2 units have the same modifiers? The best question to ask, and it only requires one figure, is a single troop-terrain combo that has a zero modifier bonus (ie value of 1). |
Posted By: Faldrin
Date Posted: 07 Feb 2011 at 10:42
|
When you clcik the XML report you see a combat terrain type: <location> <X>354</X> <Y>-322</Y> <terrainspecifictype id="25">Bleak Mountains</terrainspecifictype> <terraincombattype id="1">Large Mountain</terraincombattype>
I
------------- 
|
Posted By: Faldrin
Date Posted: 07 Feb 2011 at 10:55
|
Anyone been doing any terrain testing ?
I have started some and have some results. Ex. infantry gets -10% "bonus" for attacking small mountains (ex Mountains). -------------
|
Posted By: Mandarins31
Date Posted: 07 Feb 2011 at 14:33
|
i did some pre-tests with a friend, but i stopped after the city move. we found some bonus on plains, supposing that infantry had 0 bonus. i dont know how did you find -10% bonus attack for infantry on small mountains, because normally you can just say that you had 10% bonus difference between infantry's attack and defender's defence. you can just define a difference of bonus between the attack of an unit and the def of an other unit... and if you want to put "real bonus" you just have to chose 1 type of unit to have 0% bonus or whatever... and then you find the bonus of the other units by comparaison. but just knowing the difference of bonus between each units on a square, you can create a calculator that would give you the exact difference of strengh between an attacking and a defending army on the chosen square. it could say: attackers have 60% more strengh than defenders, for exemple. but personnally i dont know how to create that kind of calculator, and there are too many tests to do, and Gm's said that new terrain types will appear... so i stopped the tests for the moment. |
Posted By: Hora
Date Posted: 07 Feb 2011 at 14:53
|
hmm... that would be many variables, as for each terrain you would have influence on units, commanders, commander stats, etc. And you need at least to have one battle report for each variable (damn math  ) to solve that equitation system, perhaps 128 (4 terains*32 unittypes) without siegeweapons and without commander influence, so many, many battles for exact battle equitation (hmm ... if I'm not totally wrong 128*(128^n) , where n is the number of commander stats influencing battle).  Oh, I even forgot to add those NPC and special units... I would help running some armies into another for getting at least that 128 with same commander (not lvled up!), but in the end we can only hope for the GM's telling us  |
Posted By: Hora
Date Posted: 07 Feb 2011 at 14:56
|
oh, OK, n is a bit smaller (or vanishes), as we know exact numbers there... so only those 128*128, when someone is smart enough to calculate back all those commander stats... (might be some work...) |
Posted By: Faldrin
Date Posted: 07 Feb 2011 at 19:03
|
The solution is to use level 0 commanders and pure armies. Also make the assumption that troops not mentioned in the "summery" is not influenced by that kind of terrain. Yes it is many tests that is why I'm asking if anyone else is doing any testing. The tests are really easy to do. You should know either the attacking or defending bonus or lack of the same. Then you compare the result with the "expected" and add the "bonus" to the equation so the actual result equals the "expected". I have found that the bonus is different for attacking with a unit type and defending with a unit type. -------------
|
Posted By: Mandarins31
Date Posted: 08 Feb 2011 at 08:46
|
to reduce the number of tests needed per square, this is important to do pre-tests. pre-tests to know: - if attack and defence bonus on a terrain is the same for 1 unit (if i remember, i tested this, and like u i found that the bonus is not the same) - if simple and advanced units have the same atk/def bonuses on a terrain (didnt test that but i guess they have the same bonus) - if each race has different bonuses for the same type of unit on a terrain type (i didnt test, but i hope there is no racial bonus, ifnot you need to test all the units for each race) pre-tests are very important, because depending on what we could learn by these pre-tests, we could need 1024 battles per terrain, or only 6 battles/terrain. exemple1 : if different bonus for def/atk, different bonus for simple/advanced unit, different bonus for each race: 2 bonus/unit--> *2=4 bonus/ unit type (spearmen, archers..) --> *4=16 bonus/race (without siegeengines) --> *4= 64 total bonus to find --> 64/2=32=number of units to test--> 32*32= 1024 tests/ terrain exemple2 (the one i right now think is the good one): if different bonus for def/atk, same bonus for simple/advanced unit, same bonus for each race: 2 bonus/unit--> *4=8 total bonus to find (no matter which race)-->8/2=4 units to test--> 4*4 = 16 battles/terrain |
Posted By: Faldrin
Date Posted: 08 Feb 2011 at 17:20
|
I have found indications for that t1/t2 do not have the same bonuses. I will test a bit more. Units do NOT have the same attack and defence bonus on a terrain type. I don't think the they have implemented race different bonuses (yet). -------------
|
Posted By: Nesse
Date Posted: 16 Jul 2011 at 16:37
|
Building: The close quarter hand-to-hand combat required when fighting in a building is a blessing to the well trained spear or swordsman. Cavalry, however, find it extremely hard to fight here. Spear and swordsmen find that building fighting is their place to shine. An intimate knowledge of the choke points in a building also somehwat benefits defending bowmen despite their limited arcs of fire. Cavalry detest fighting in and around buildings. |
Posted By: Nesse
Date Posted: 16 Jul 2011 at 17:36
|
Mountainous area: Even the smaller crags in this mountainous area will hinder attacking cavalry, as well as more heavily armoured swordsmen. Unlike cavalry - who suffer from the difficulties in mustering a coherent defence in mountainous terrain - defending ranged units find every advantage in the high lookouts and rocky ground. |
Posted By: Nesse
Date Posted: 16 Jul 2011 at 17:37
|
Heavily wooded terrai: Heavily wooded terrain spells difficulty for ranged units and cavalry. Infantry units find they have a large advantage here. Like the attackers, these thick forests do not provide much assistance for anyone except your footsoldiers, who take great pleasure in the extra cover afforded to them. |
Posted By: HonoredMule
Date Posted: 16 Jul 2011 at 17:40
Early in the game, at least, there was definitely no difference between T1 and T2 for terrain bonuses. However, the bonuses have always been unique per combination of unit type, terrain, and position (attacking or defending). So you would have to be very careful to eliminate all of these dependent variables before accurately identifying a new one. And unfortunately, there is some random factor, which means there's a variable dependent that cannot be eliminated and just be measured statistically over multiple trials before it can be separated from any other unknown factors. |
Posted By: Nesse
Date Posted: 16 Jul 2011 at 19:25
|
For attacking units in mountainous area "Even the smaller crags in this mountainous area will hinder attacking cavalry, as well as more heavily armoured swordsmen." This seems to imply that heavy infantry is different from light infantry - in mountainous area. Probably no difference in most places... |
Posted By: HonoredMule
Date Posted: 16 Jul 2011 at 19:43
| I wouldn't read too much into that. I believe the phrase is just describing swordsmen in general, contrasted against cavalry. |
Posted By: Mandarins31
Date Posted: 16 Jul 2011 at 21:26
Well, we can admit this random factor is not that important. if this change something like a maximum of 5% this is not that much and we have anyway a precise idea of the difference of terrain bonus for a batle. Also we could apreciate the amplitude of this random factor by doing the same battle few times. |
Posted By: Nesse
Date Posted: 16 Jul 2011 at 21:51
|
Open plains: (And here it is light spear that has special treatment) Cavalry comes into its own when able to strike hostile forces at will, and from unexpected directions - and nowhere is this more feasible than on open plains. Lightly armoured spear units, however, prefer terrain where there's some cover available. Fighting defensively on open plains, cavalry draws strength from the ability to form and reform their lines of engagement depending on the direction of battle, and it is here where cavalry excels. |
Posted By: Nesse
Date Posted: 19 Jul 2011 at 10:47
|
Very mountainous terrain: Swordsmen, and especially cavalry, find attacking into very mountainous terrain difficult - although ranged units can excel. Narrow passes and gulleys hinder any mounted units' ability to defend adequately. Nimble spear units have no such worries here, and defending bowmen are able to use their height advantage to maximum effect. |
Posted By: Nesse
Date Posted: 19 Jul 2011 at 11:03
|
Small hill: Attacking up a small hill offers little advantage, or penalty, to the attacking force. Defences with a height advantage, however small, are better than nothing. |
Posted By: Nesse
Date Posted: 21 Jul 2011 at 09:25
|
March (yes, march counts as large hills...) Large hills such as this are difficult for attacking mounted units, but provide opportunities for ranged to unit to show their worth. Arms-bearing footsoldiers take it all in their stride, of course. Manueverability and lines of sight are the keys to defending very hilly terrain. Lightly armoured units such as spearmen and ranged units find it easy going here. |