OOS Combat Resolution (Was: OOS Missile firing test.)

wyvern11 · Post by **wyvern11** » Fri, 29. Jun 07, 08:22

@jlehtone & googlemeyer

after your questions i did a test if there are differences between
- watching sectormap
- not watching sectormap (only ocassionally watching logbook)

i took hitpoint-log as well to see damage dealt easily and concerning ttls remark what may or may not influence OOS-combat i made some really really strange observations, --- for those of you who do not want to know exactly, on which silken thread the fate of their ships, crew and credits hang, i would have put the info in spoilers but "code" is always yellow

i will give the results without comment or opinion - read it and think for yourself

EDIT: Table may have systematic error - details see next post

Code: Select all

percentage of player wins (40 replays):
0,45 = 45%

             |  1280x768       |  1280x768        |   1024x768      |   1024x768       
             |  fullscreen     |  windowed *      |   windowed      |   fullscreen         
---------------------------------------------------------------------------------------
map          |                 |                  |                 |  
watched      |      0,45       |      0,22        |     0,425       |     0,46
             |                 |                  |                 |  
---------------------------------------------------------------------------------------
map not      |                 |                  |                 |  
watched      |      0,35 (0,25)|      0,15 (!)    |     0,25        |     0,2 
             |                 |                  |                 |         		
---------------------------------------------------------------------------------------

* for some reason only got botched windowed mode

overall sucess rate when watching 0,39
when not watching 0,23 (160 replays each)

info about my machine: ath64 3200+ 1GB ram x800GTO

this was under a beginning player, argon trader out of the box, combat scripted - for my regular save, upper table contains rate in (...)

we'll see which info can be gotten out of hitpoints-tables in a while

jlehtone · Post by **jlehtone** » Fri, 29. Jun 07, 09:11

wyvern11 wrote:this was under a beginning player, argon trader out of the box,

That is good. Rank, playtime, and property are all near 0, so their influence is counted out. I was about to suggest that setup, but forgot. What about using Xenon L vs identical player L as the combatants in one OOS sector and then adding player M2's to second OOS sector? The combatant (M3) would not add player property significantly, and M2's with no commands do not run any scripts nor need rendering, thus not add much CPU load. That ought to show if the count of player M2's does count.

Those hitpoint-tables ought to reveal (easily) how often attacks are resolved in each case. Average damage per attack and variation of that damage are there too, but need more number crunching.

Your efforts are much appreciated, wyvern11.

Edit: Drones. Fighter Drones. Traders do launch them when attacked, both in sector and OOS. Is that limited to TS/TP? And does it even matter as the distraction may be too little and too late?

wyvern11 · Post by **wyvern11** » Fri, 29. Jun 07, 10:51

xenon L ? maybe

if you can provide me with a rough estimate, how long the fight will take - or if you are interested to know hitpoints or not (i think this will be a one shot one kill situation)

any idea of an unknown sector, where spawn of not-so-innocent-bystanders is most unlikely?

in my real combats i use disco's as drone-replacements (although it is difficult to measure the effect i believe that there is even OOS the outcome better because J/K might attack a disco instead of M2 which it will most definitely kill in one shot)

EDIT !!!! Important - i found a clue in the list of hits, why non-watching makes a difference:
at non-watching
- nothing happens during first 30sec
- since i break off at 40sec fight may not be over then
- because xenon will allways be on position two, it will be overrepresented

correspondingly having different resos not matching optimal combat may become even more laggy -> even more xenon overreps!!

EDIT2: hitlist looks like
- hints for damage taken sort of like grouped in bands - difficult to sort out
better take a ship with single battery / front / back mounted wapon only and large shields - transport for example

another Edit: from the damage in non-watch situations it looks like it isn't higher but as if similar combat occurs, only six times slower

wyvern11 · Post by **wyvern11** » Fri, 29. Jun 07, 19:27

rather than making the next edit i'll open a new post

- it was as i feared - forget my table between watched and non-watched fights

both rates are fairly equal - my fault was, that in some cases (nonwatching) i ended the battle when both ships were still alive - counting top down, xenon was always last in map - so xenon was awarded the win incorrectly

so 0,45 is the rate for a beginning player in OOS combat whatsoever

as i thought watched an unwatched are fairly the same - nonwatched yields same damage, but 6 times slower (5sec : 30 sec). main difference is shield recharge is watched and unwatched every 3sec

hitpoints on a larger number of cases now rather resembles an uniform distribution an not the expected bands - details tomorrow

i will provide some examples tomorrow

jlehtone · Post by **jlehtone** » Fri, 29. Jun 07, 19:47

wyvern11 wrote:as i thought watched an unwatched are fairly the same - nonwatched yields same damage, but 6 times slower (5sec : 30 sec).

Interesting, same damage per hit, regardless how often you shoot? Kind of logical, but if "hit" would represent result of combat from a period of fighting, then it is not so logical.

Dragon, only AIRE in turret, 100% on Thunderbolts. This time I watched the sector map and timed the disappearance of the Pirate Base: 5:39. 67 missiles were consumed in that time. Damage value of Thunderbolt is 75k. Thus, 60 of them would consume all of original shield and hull, if they do full damage. Base can recharge about 251 MJ in that time, which would require four additional full strength missiles. Since three more than that was consumed, the missile damage is close, but not quite the nominal. So a missile every 5 seconds for watching the map.

Raptor, now 4 AHEPT and 2 BPAC, one in each turret. Looking at Base, the shield drops every 30 seconds, so turrets do not seem to fire independently. Shield down in 5:41. Bit slower than with 6 AHEPT, as expected.

googlemeier · Post by **googlemeier** » Sat, 30. Jun 07, 00:43

Meantime the fleet has only 7 and 6 K left, the survivors fired up to 5 hornets each - firing set to 80%.
Added torpedoes and changed to 100%.

slowly I run out of Ks...

eladan · Post by **eladan** » Sat, 30. Jun 07, 04:42

Very interesting stuff indeed.

@wyvern, are you setting fightskill on the ships? IIRC, ships are given a random value for fightskill if one isn't set.

I'm rather interested whether it has any effect OOS - the things it supposedly affects would be more useful for IS combat - attack run length, dodging ability, etc. (though it's supposed to affect missile usage as well.)

wyvern11 · Post by **wyvern11** » Sat, 30. Jun 07, 07:18

no - did not take fightskill into account.
found the command to change it but had no idea which value to assign
do you ?

in my regular game i did not notice that long running patrols performed better than fresh ones though

eladan · Post by **eladan** » Sat, 30. Jun 07, 07:50

It's used in the combat scripts plugin.acp.xxx... where the maximum FS used is approx. 24 (ish - can't recall exactly, and don't have the files to hand.) There are 7 discrete levels used, where it takes a FS range and applies fight values. E.g. at the lowest setting, with FS=0-2(?), it's speed is roughly half the max speed for that type of ship, and has a very low (5%?) missile chance, amongst other settings. At the highest setting with FS~24 speed is maxed out or close to it, and missile chance is in the upper range as well.

I haven't done enough testing, either IS or OOS to see how much of a difference it makes, but have been told it works, at least IS. I very much doubt those scripts are used OOS, but the game engine may well use FS in its own OOS calculations. In that case though, I have no idea what the maximum value it uses would be. If testing it, I'd probably try FS=100.

jlehtone · Post by **jlehtone** » Sat, 30. Jun 07, 09:27

fight.attack.object calls plugin.acp.fight.attack.object only if this ship is Fighter, Drone, M6, or M7. Never with M1 or M2. And skill is used only on plugin.acp.*. Thus, M1 and M2 behaviour is not affected by skill, not even in sector.

NPC ships can dynamically adjust their missile firing probability in certain situations. The base chance is defined by the skill. About 95% of claimed Xenon fighters have missile probability at 0%. But some ships have it even as high as 100%.

wyvern11 · Post by **wyvern11** » Sat, 30. Jun 07, 10:16

any ideas to my previous questions about
- a rather spawn-free, empty unknown sector? although i scrap every ship in sector before each replay, there is a chance of respawns during testing

- which ship to test instead of colossus
(single gun bay, low damage high shields maybe, non capital because of minimal fightrank/fleetsize)?

@jlehtone

Interesting, same damage per hit, regardless how often you shoot?

i think it's one shot/hit per round - only round is longer
to my actual opinion fight is very much the same if you watch or not - only if you watch it is 6 times faster

unnoticed due to testing and playing i lost a raptor with 40 LX on board - you know what ? i didn't even care

jlehtone · Post by **jlehtone** » Sat, 30. Jun 07, 10:33

wyvern11 wrote:any ideas to my previous questions about
- a rather spawn-free, empty unknown sector? although i scrap every ship in sector before each replay, there is a chance of respawns during testing

- which ship to test instead of colossus
(single gun bay, low damage high shields maybe, non capital because of minimal fightrank/fleetsize)?

Aren't there couple sectors off-map, used only in plot? If there are, is there any way to access them?

How about Teladi Falcon Sentinel. M3, 250 MJ shield. It has turret, but if you load it up without turret gun. Say 8 AIRE?

Some might consider 40 LX "valuable". They might learn to not worry.

eladan · Post by **eladan** » Sat, 30. Jun 07, 10:47

jlehtone wrote:fight.attack.object calls plugin.acp.fight.attack.object only if this ship is Fighter, Drone, M6, or M7. Never with M1 or M2. And skill is used only on plugin.acp.*. Thus, M1 and M2 behaviour is not affected by skill, not even in sector.

Ah - thanks jlehtone, been a while since I checked the scripts.

Still, possibly worth checking, if only to confirm that there's no hardcoded dependency in OOS combat.

wyvern11 · Post by **wyvern11** » Sat, 30. Jun 07, 13:16

here is an excerpt of hit-list of non-watched combat:

Code: Select all

N    t(s)      ship    hitpoints taken
0	32	 Xenon Colossus 	1428132
5	32	 Ihr Colossus 	1487128
10	32	 Ihr Colossus 	1597133
17	32	 Xenon Colossus 	2229240
17	62	 Ihr Colossus 	2253165
18	32	 Xenon Colossus 	2593710
18	62	 Ihr Colossus 	2676183
18	92	 Ihr Colossus 	2790038
23	32	 Ihr Colossus 	2866702
24	32	 Xenon Colossus 	3302703
27	32	 Xenon Colossus 	3388650
37	32	 Xenon Colossus 	3754400
37	62	 Xenon Colossus 	4006722
38	32	 Ihr Colossus 	4320483
38	62	 Xenon Colossus 	4658835
39	32	 Ihr Colossus 	5634264

i omitted the shield-reloads every three seconds in between - the 6200000+ hits (instant deaths) too

watching looks like:

Code: Select all

34	22	 Ihr Colossus 	360344
37	12	 Xenon Colossus 	404327
37	17	 Ihr Colossus 	427047
...
39	17	 Ihr Colossus 	2233110
3	12	 Ihr Colossus 	2239529
35	7	 Ihr Colossus 	2240525
...
8	7	 Ihr Colossus 	5835060
15	12	 Ihr Colossus 	5886496

you see, there's hits every 5sec instead of every 30, but amount about the same (somehow there are two seconnds offset)

jlehtone · Post by **jlehtone** » Sat, 30. Jun 07, 14:59

Only three of those listed hits do less damage than a Torpedo (assuming a Torpedo makes 950.000 points) and average hit is way more. Thus, missiles would not help Capital ship.

The 2 second delay, I think I have seen it too. Could it be the couple 'wait' calls in the fight.attack.object before it reaches the "fire weapons" instruction? Similarly, I've seen a hit at ~13 seconds and then every 30 seconds. I assume that when I have given the attack command, exit that ship and open the Pirate Base station dialog, I have sector map open briefly. That lets "watched" hit at ~2, ~7 and ~12 seconds. From then on I'm watching the "nonwatched".

wyvern11 · Post by **wyvern11** » Sat, 30. Jun 07, 15:19

Average Hit was about 2.5mio points, thus you are correct that in a one on one situation missiles might be in worst case useless and in best case just tip the scales. in fact in a larger number of tests rate of shots below 1GJ is 0,13

but in all sensible cases player should have at least 4:1 superiority in arms thus up to four firestorms might defintely lead to another outcome

googlemeier · Post by **googlemeier** » Mon, 2. Jul 07, 00:10

Now 472 is black again - lost 10 K, 10 P, 20 LX, 40 L and four sats.
And the firestorms, the Ks were loaded with...

zazie · Post by **zazie** » Mon, 2. Jul 07, 11:12

jlehtone wrote:fight.attack.object calls plugin.acp.fight.attack.object only if this ship is Fighter, Drone, M6, or M7. Never with M1 or M2. And skill is used only on plugin.acp.*. Thus, M1 and M2 behaviour is not affected by skill, not even in sector.

Very interesting, could explain partially my observations (didn't know about the veterancy (sp?)/ fighskill ranking) - my "solid standing" patrols were rather experienced (280 or more flying hours without docking at a station, continuous/regular combat situation).

Another observation I don't know if it is known or not:
The engine seems to treat "Clusters" as own objects and does not 'switch' to Fighter/Scout or some commands are more efficient than others.

Yesterday while building a complex I sent an Orca to Kingdom End's shipyard. Weapon setup:
F: 2*bHEPT
TR: 1 aHEPT , 1 aPBE
TL: 1 aHEPT , 1 aPBE
B: 2*bHEPT

When I quickly checked the sector I saw a Kha'ak-Cluster. I gave the "Attack ..."-order to my Orca and watched on the sector map (sometimes with SETA):
The Cluster split (it was a big one) but to my surprise the Orca did not kill one single Scout though the shields went down considerably and I got the "Attention, one of your ships is under attack"-message several times. I am sure: no kill at all.

I decided then that the Orca should dock at the shipyard and was about to jump into the sector after Orca's docking to clean up the situation.
But as soon as the Orca had a new command it was instantly able to kill the Kha'ak-ships following and attacking him. The Orca killed all (at a very regular rhythm) but two that were in a dog-fight with two smaller Boron ships.

Possible explanations:
1. "Attack Cluster" does not switch to "Attack next target" : probable but not useful. The ship should defend itself.
2. Order DOES switch but "Attack next target" is highly less efficient than "Defend ship".

This last point is the very reason why I post it here. You probably know more.

jlehtone · Post by **jlehtone** » Mon, 2. Jul 07, 14:36

You gave "Attack ..". That command completes when the target dies/vanishes. Then ship will "run" 'None'. It won't react to attacks on 'None'. It might be that it does react when it runs any command. Some commands (e.g. 'Idle', 'Standby', 'Attack all') definitely do react to attacks.

The XaiCorp Group Management System script offers "default command". The leader will automatically start "Attack all enemies" when its previous command completes. Much better than 'None' in combat situation, but not necessarily optimal action in station hauler TL.

zazie · Post by **zazie** » Mon, 2. Jul 07, 17:41

OK thanks.

So it does not give any input to the topics of this thread - sorry.

egosoft.com

egosoft.com

OOS Combat Resolution (Was: OOS Missile firing test.)