Looting tested

正在查看此主题的用户

Sicnarf

Regular
How the looting skill affects how many items you get and their quality.
To read the conclusions just jump to the bottom, everything before that is how the testing was done.

In case you missed it last time -> Map speed/pack animals, Trade skill and town relations tested
http://forums.taleworlds.com/index.php/topic,46362.0.html

Setup.
All tests were conducted using a cheated character with 30 in all stats and 10 points in all skills bar looting. I first collected 94 units, 5 companions and myself. I then cheat teleported as close as I could to a group of 30 Rhodok trained spearmen deserters (From now on referred to as spearmen) the largest even numbered nutral part I could find on the map. At this point I saved the game and made a backup of the save.

Testing.
I then disbanded all my troops and told my companions to leave so I was alone, I then saved the game and proceeded to attack the spearmen and cheat killed them with Ctrl+Shift+F4, took all the loot they dropped and sold it at Dhrim to the armour merchant while noting down the following. How many items were dropped, how much money I got at the end of the battle, how much experience I gained at the end of the battle and how much all the items sold for. I then reloaded the game and did the same again 9 more times before averaging the results and putting them in the table below. I then attacked and manually killed the spearmen on easy, then difficult and ensured that the results fell within the margin of error for the previous 10 runs to ensure that cheat killing or difficulty did not skew the loot given.
I then reloaded the game added 5 points to looting (+2 leader bonus) and repeated the above, then added 5 more points for 10(+4) looting and repeated the above.

I then restored the save game from backup and disbanded all but 14 units for a group total of 15 and repeated all the above again.
I then restored the same game from backup and kept all 94 units + 5 companions for a group total of 100 and repeated all the above again.
I then restored the same game from backup and disbanded all but 49 units for a group total of 50 and repeated all the above again.

*Note, After doing the 100 unit test I stopped noting the after battle Denar and Experienced gained as it was clear by then that lotting had no effect on them.

All items were sold in Dhirm to the armour merchant, my cheated character had 10 trading and 99 relation with the city.
All 144 test attacks on the spearmen took place within seconds game time in the hope of limiting and randomness that may occur due to time passing in game. However despite how close I was the spearmen sometimes turned away or charged towards me and I ended up fighting them on maps that differed from one another.
When averaged all numbers bar Number of items dropped were rounded down.
Trying to keep track of Item prefixes was to time consuming which is why I opted to sell the loot and use the sale price for testing.
The quantity and quality of loot received differed by a large margin within all ten 10 runs for any given looting skill. For example solo with loot 0, the least items found was 12 and the highest was 23, best selling loot was 1681 and least selling was 753. The averages received were a good meter stick but for very accurate average values you would need at least 50 rounds for each test which their is no way in hell I'm doing 624 battles even with cheats. So do bear in mind for the conclusions that the data is not optimal.

Legend:
Num = Number of items dropped.
Sell = How much all the loot sold for in Dhirm.
Den = How much after battle Denars were received.
Exp = How much after battle Experience my character received.
AveItemValue = Average value of each item sold.

Test 1
----------Solo character----------
Looting level 0,
Num Sell Den Exp AveItemValue
17.1 1233 1235 1307 72

Looting level 5+2, Solo character.
Num Sell Den Exp AveItemValue
16.2 1310 1187 1311 80

Looting level 10+4, Solo character.
Num Sell Den Exp AveItemValue
15.6 1149 1218 1167 73

Test 2
----------Me + 14 unitss----------
Looting level 0,
Num Sell Den Exp AveItemValue
14.4 911 551 563 63

Looting level 5+2,
Num Sell Den Exp AveItemValue
18.4 1325 519 560 72

Looting level 10+4,
Num Sell Den Exp AveItemValue
16.4 1211 548 510 73

Test 3
----------Me + 49 units ----------
Looting level 0,
Num Sell Den Exp AveItemValue
6.9 452 65.5

Looting level 5+2,
Num Sell Den Exp AveItemValue
10.5 812 77.3

Looting level 10+4,
Num Sell Den Exp AveItemValue
15.9 983 61.8

Test 4
----------Me + 5 companions + 94 units ----------
Looting level 0,
Num Sell Den Exp AveItemValue
4.1 323 105 104 78

Looting level 5+2,
Num Sell Den Exp AveItemValue
7.1 564 104 108 79

Looting level 10+4,
Num Sell Den Exp AveItemValue
9.9 661 103 97 66.7

CONCLUSIONS:
Ok easy one's first:
1)Looting appears to have no effect on the after battle experience gained by your character, oddly though according to the testing at 10(+4) looting you always averaged less experience gained than other levels. But I think I can safely say that is just a concidence.
2)Looting appears to have no effect on the after battle Denar gained.

Ok now the hard one's:
It would appear that looting does not actually provide you with more loot at all according to test 1. Looting just counteracts the theiving of your troops so you end up with more of the loot instead of them. If you are going solo looting does nothing for you, once you start adding troops and companions to your group you begin to receive less of the loot as they are getting a small share for themselves.

With 0 looting ability you get the following drop in loot as you add units.
15 = 25% less loot
50 = 65% less loot
100 = 74% less loot

With 5(+2) looting ability.
15 = 0% less loot (Actually -7% but I doubt their was an increase)
50 = 35% less loot
100 = 45% less loot

With 10(+4) looting ability.
15 = 0% less loot
50 = 21% less loot
100 = 47% less loot

Moral of the story is that for the best loot keep your party as small as you can, going to 10(+4) appears to be a waste of time as the biggest gains to be made for the least point investment is 5(+2), you would be better off getting a companion to mule this skill as maxing it gives very small gains, gains that are easily overshadowed by the trade skill. Average Item values don't change that much at all so the chances of getting lordly items does not appear to be related to looting, if your going to get a lordly item you will, looting just helps cut down on the odds that one of your troops will get to it first.
Please bear in mind that all that stats above were averaged over 10 battles so their accuracy is not close to perfect so their are bound to be differences between what I averaged and what is 100% truth. So take the above with a grain of salt. The only thing I'm 100% certain of is that looting does not actually generate more loot, it just counteracts your troops taking stuff for themselves.

Edit : Someone posted that village looting time is also affected which got me thinking that loot gained might also be effected, so back to the testing character. Long story short when you loot a village you get the same amount of loot on average no matter what level of looting you have or how many troops you have with you.

Setup
Same character as above with 99 men, looting starting at 0 then incremented to 5(+2) then 10(+4), only averaged 5 runs for each test as all looting results were very consistent. All items sold in Dhirm, due to the shops limited space I always sold the first 30 items to the armourer and the remainder to the Weapons smith, all items were sold on the same day to avoid price fluctations.

Legend:
Num = Average number of items looted.
Sell = How much all the items sold for.
Den = How much denar I looted.
Catt = Average number of cattle looted.

Testing
----------Me + 99 units ----------
Looting level 0,
Num  Sell   Den  Catt
37.2  3977   251  3.6

Looting level 5(+2)
Num     Sell Den   Catt
37.8     4182 216   3

Looting level 10(+4)
Num     Sell Den   Catt
36.4     4245 351   2.4

I then disbanded the 99 men, reduced looting to 0 and started again, after 3 test runs it was very clear that their was no change in the average quantity or quality of loot even with my character doing it solo. For village looting the looting stat only reduces time taken and additionaly it does not matter how many men you have with you they do not take any loot for themselves.
 
Cool that your doing all of these things! Nice to get some answers to the mysteries of calridia
 
So, I may be jumping to conclusions, but according to your data I would say the "looting" skill is either broken or useless. Getting looting to 5 even with a companion doesn't seem worthwhile to me. Can someone fix this? *crosses fingers that TheMageLord will show up soon*
 
theguruofreason 说:
So, I may be jumping to conclusions, but according to your data I would say the "looting" skill is either broken or useless. Getting looting to 5 even with a companion doesn't seem worthwhile to me. Can someone fix this? *crosses fingers that TheMageLord will show up soon*

It also lowers the time to it takes to loot a village or steal cattle, which can be great if you're going to play the game that way.
 
I would suggest running each scenario about 5 times, and see if there is much variance. Maybe, for example, with 10(+4) Looting skill you just got a low-end random amount in the single test run, and could possibly have had something like:

Looting level 10+4, Solo character.
Num  Sell  Den  Exp  AveItemValue
22.0  1500  1400  N/A  (whatever)

In other words, more tests per case to get a better average.
 
bravesirknight 说:
theguruofreason 说:
So, I may be jumping to conclusions, but according to your data I would say the "looting" skill is either broken or useless. Getting looting to 5 even with a companion doesn't seem worthwhile to me. Can someone fix this? *crosses fingers that TheMageLord will show up soon*

It also lowers the time to it takes to loot a village or steal cattle, which can be great if you're going to play the game that way.

I don't so I've never noticed. I take the honorable path and don't steal from villagers.


Oh, and if you have to run more tests, use the CTRL+SHIFT+F4 cheat so you kill them faster...
 
bravesirknight 说:
It also lowers the time to it takes to loot a village or steal cattle, which can be great if you're going to play the game that way.

Forgot about that as I stopped looting vilages a long time ago, did some testing and looting has no effect on the loot gained from villages, nor does the number of men for that matter. Updated my post with the new info.

HardCode 说:
I would suggest running each scenario about 5 times, and see if there is much variance. Maybe, for example, with 10(+4) Looting skill you just got a low-end random amount in the single test run, and could possibly have had something like:

Looting level 10+4, Solo character.
Num  Sell  Den  Exp  AveItemValue
22.0  1500  1400  N/A  (whatever)

In other words, more tests per case to get a better average.

I ran each test 10 times then averaged the results, it's mentioned under the testing section of my post.

allthesedamnnamesaretaken 说:
Oh, and if you have to run more tests, use the CTRL+SHIFT+F4 cheat so you kill them faster...

I did use CTRL+SHIFT+F4 for 120 of the test rounds, I only killed them manually 24 times to ensure the cheat was not skewing the loot results, it's mentioned under the testing section.
 
Im hearing that basically looting is useful at a low level for when you have lots of people...

thanks for doing the research, saves me the skill points :smile:
 
Very well done with lots of patience and determination shown to go through such a detailed procees. I'm glad that you've done this, because I knew I was taking a hit on loot, but didn't know how much it was.

Maybe you should run it a few more times to get enough data for an ANOVA, and a couple of t-tests!  :razz:

Although, if you are up to it, I'd be curious to know the results from an army of even larger size, say 150.
 
Wraithcat 说:
Very well done with lots of patience and determination shown to go through such a detailed procees. I'm glad that you've done this, because I knew I was taking a hit on loot, but didn't know how much it was.

Maybe you should run it a few more times to get enough data for an ANOVA, and a couple of t-tests!  :razz:

Although, if you are up to it, I'd be curious to know the results from an army of even larger size, say 150.

Not a chance that was very boring work and interfered with my playing of the game I don't want to go back, I'm going to do tactics now but the little testing I have done has provided results that are not really matching up with what the description says.

Cheated characters party size only goes up to 103 but thats not really an issue I just find that my army size never goes beyond 125 before I'v finished taking over the map for my side so 150 size tests would be kind of pointless. Anyway once you hit 100 loot becomes less of an issue due to having a few towns for money generation.
 
Looking at the scripts, looting is added into the calculation for 3 things:

1. Cattle stealing, it adds to the max amount of cattle you can steal, 2 per looting skill.

2. Factors into :loot_probability. I'm not entirely sure how the troop_loot_troop function works, but judging by the name :loot_probability I can only assume it's some kind of chance of getting items. How this is calculated is it takes the looting skill + 10 multiplied into :loot_probability and then divided by 10. So it should be adding 10% to this number per looting skill. Then it divides that by :num_player_party_shares, which is bigger if you have a bigger party.

3. Goes into the raid progress for raiding villages. Default raid progress is 8, plus 1 for every 2 points of looting skill. So that's adding only 12.5% for every 2 points, not counting odd numbers of looting. So 10 looting would make it progress 62.5% faster.

It looks to me like it's working fine, looking at the numbers you obviously have significantly more loot with more looting skill except on a few cases. As random as getting loot is, someone would have to do a LOT more tests to get a defininitive number on looting. The only easily available definite number is looking at the code, which is adding 10% chance per level, which is exactly what the description says. You might even be getting more loot at the solo level, you just didn't get lucky on those tests. Either that or theres a maximum benefit level for :loot_probability to be at, but even then it would be very helpful when you add more troops - so still worth it in my opinion.
 
Well I said that I only tested 10 rounds for each which is not perfect as loot varies so much, but their is definatly a cap on how much loot you can get, in no test at any level did the amount of loot gained go more than a little over the solo average and at 10(+4) I should have been getting 140% more loot and nothing ever came close to that. I figure that you can raise loot slowly because untill you hit 50+ troops it's not really helping to much anyway. Thier is no way I'm going over 5-6 loot on a companion because by the time I hit 100 men my income is mostly from towns anyway.
 
后退
顶部 底部