Topic: Armor Type relative strengths

[edxs]

I ran into Ecthelion's combat tester, so I decided to run a bunch of tests (don't ask how long it took) to determine the relative strengths of the armor types.

Assumptions:
2 combatants, each with 1000 hits, +100 OB, no DB, 10000 trials of each combination.

I plan to do it with +30 DB to simulate fast PCs, but it will take another long while. If I did it again I would assume +50 or +75 OB to simulate reasonable parrying (expected roll with +100 is 150, which ignores some of the easier to-hit targets at high ATs) and do more trials.

Results:
20   0.74675
19   0.66425
18   0.615
16   0.60575
15   0.59475
17   0.50825
14   0.501
12   0.49375
13   0.47325
11   0.4685
8   0.459
10   0.45675
7   0.44075
9   0.4365
6   0.43275
4   0.43225
3   0.43025
5   0.4225
1   0.412
2   0.406

It seems that rigid leather is fairly bad for melee types, though the reduced missile penalty is good for ranged weapons. It was nice to see that being naked is not, in fact, better than light armor.

[markc]

 Nice stats and analysis.
MDC

[Emaughan]

Great idea (you got an idea point).

My question is - how did you factor crits, or is this only based on concusion hits?  Which weapons did you use?  Is this a compilation of all arms law weapons, or just some, or only one?  I'm not up to speed on Ecthelion's combat tester and the not sure the parameters it involves.

The greatest critism that I have with Arms Law is that the crits start sooner for many of the armors before naked folks!  Sure the crits and concussion hits taken by naked man get nasty high up in the charts, but those of who have played a lot know that any crit can spead up the demise of the reciever of that crit.  One round of stun can totally shift momentum in combat (as it should).  That is why I feel it is worse to were armors that increase the range in which one can recieve any crit than going without any armor at all.  Combat campanion did much to fix this shortcoming, but I miss the greater variety of the older charts and A-E crits.  I hope ICE takes the best of CC (armor by the piece, better protection from armor, and one skill cost for all manuver in armor for each profession) and mix it with the best of arms law (specific weapon charts, more variety with A-E crits).

[edxs]

My question is - how did you factor crits, or is this only based on concusion hits?  Which weapons did you use?  Is this a compilation of all arms law weapons, or just some, or only one?  I'm not up to speed on Ecthelion's combat tester and the not sure the parameters it involves.

I don't know much about the guts, but it seems to resolve crits properly, with at least stuns and bleeding. I used broadsword for both combatants.

I've started on testing with 50 parry and +30 base DB, and the results are changing pretty considerably. The first post is certainly applicable if you're getting waled on (constant 100-150 rolls), but not so much for battle of attrition where the bottom half of the chart comes into play.

[edxs]

Here it is:

AT   Win%
20   0.6347515
18   0.579099
19   0.5635215
16   0.56286
17   0.5518645
15   0.5467905
14   0.545583
13   0.531085
4   0.5122825
3   0.49948
10   0.4987805
12   0.495171
9   0.47787
6   0.4703285
11   0.4603035
1   0.457173
2   0.4239345
5   0.4075
8   0.392131
7   0.3894905

I know I'm staying clear of AT5, 8, and 7. I'm not really sure what makes 6 so superior. AT 4 (and 3) is incredible, and I'd be extremely hesitant about ever giving it out.

[rdanhenry]

AT 3 and 4 aren't supposed to be available as wearable armor (normally). AT 5-8 are more for protection against environmental hazards than actually attacks. Looks like the actual harden leather armors do have an advantage over AT 1, although not a very great one. Of course, it is very easy to tweak this for a particular campaign by varying the availability of different item types. A common +10 non-magical leather armor supply (faerie cattle + elvish craftsmanship) can help make the light armors more attractive.

[OLF, i.e. Olf Le Fol]

ATs 3,4, 11 and 12 aren't supposed to be wearable armours. As for ATs 5-8, they're "heavy outer garments normally worn as weather protection by certain civilians and as combat protection by some militia and irregulars." As such, I'd still expect them to be better than normal clothes (and no armour at all), which doesn't seem to be the case if you have a high DB, due to high quickness (if I read both charts correctly).
So, the best non-metallic armours would be AT10? Interesting to know~

[kevinmccollum]

Run that simulation with Claws or other animal attacks instead of weapons. You'll see where some of the soft leathers actually have a benefit.

[Ecthelion]

My question is - how did you factor crits, or is this only based on concusion hits?  Which weapons did you use?  Is this a compilation of all arms law weapons, or just some, or only one?  I'm not up to speed on Ecthelion's combat tester and the not sure the parameters it involves.

The simulator uses formulas to more or less accurately represent the Arms Law tables. For the criticals it uses the Middle Earth RPG critical tables, which are a simplified form of the RM crits.

[Ecthelion]

Assumptions:
2 combatants, each with 1000 hits, +100 OB, no DB, 10000 trials of each combination.

I used different combatants: 80 hits, +80 OB with Scimitar, 15 DB, 25 shield, 10000 trials. Combatant B always uses AT 1, combatant A uses each of the ATs 1 to 20. Here are the results:

AT / percentage wins of combatant A
20 0.89
19 0.84
16 0.80
15 0.79
18 0.78
14 0.73
17 0.68
12 0.69
13 0.67
11 0.65
10 0.58
4 0.56
3 0.56
8 0.54
9 0.52
7 0.50
1 0.49
6 0.48
5 0.44
2 0.43

It looks a bit different from your results. The main reason for this is probably the very high number of hit points you assigned. This IMO makes the concussion hits (and also bleeding) given from criticals and the attack tables an almost irrelevant factor - which should not be the case.

But also my above results should be taken with a grain of salt. They might differ depending on the weapon used (as already mentioned above, the animal attacks might give quite different results), the OB/DB/parry ratios etc. A factor that I also did not consider above is the Qu penalty. Many of the higher ATs would fare far worse if the DB of the combatant was reduced due to the inherent Qu penalty of the armor.

[OLF, i.e. Olf Le Fol]

When trying the program, I have a "java.lang.ArrayIndexOutOfBoundsException: 8" error. ^^;;;;;;;
Any possibility to have the source code, so that I can debug and compile myself?

[Ecthelion]

When trying the program, I have a "java.lang.ArrayIndexOutOfBoundsException: 8" error. ^^;;;;;;;
Any possibility to have the source code, so that I can debug and compile myself?

I won't hand out the source code, sorry. But you may try to call the program from the command line via java -jar <program-name.jar> to see a more detailed debug message (if possible).

[OLF, i.e. Olf Le Fol]

Never mind: I found it. The system just doesn't like it when a weapon is both Slaying and adds an additional critical. ^^;;;;;

[edxs]

I think this approach has some merit, but I'm thinking that perhaps the best approach would be to use the AL tables and generate expected damage and critical severity for various allowed OBs. For example, 0, +5, +10, +25, +50, +75, +100. Bersekers against high level foes would look at the +100 row, high quickness parrying fighters or low level PCs towards the bottom of the chart.

This would be easy enough except for the data entry involved. What would be really useful is expected bleeding, stun, parry/no-parry, and injuries, but frankly I don't know if anyone has enough patience for all that.

[edxs]

ATs 3,4, 11 and 12 aren't supposed to be wearable armours. As for ATs 5-8, they're "heavy outer garments normally worn as weather protection by certain civilians and as combat protection by some militia and irregulars."

I was under the impression that 3,4,11 & 12 were not normally wearable, but the armor penalty table rows existed in the case that they were available as magic armor or whatever. Animals don't get quickness/movement penalties for their skin, for example.

[edxs]

While this approach was certainly less time consuming, even considering the scripting time, the results just aren't as satisfying. They're certainly useful, though. I dislike not having an obvious ranking for a given situation.

It looks a bit different from your results. The main reason for this is probably the very high number of hit points you assigned. This IMO makes the concussion hits (and also bleeding) given from criticals and the attack tables an almost irrelevant factor - which should not be the case.

Initially I wanted to be sure the combat was lasting a long time to even out the results, but I suppose I forgot about all those "Die in 6 rounds" and "target dead" results, which would certainly dominate the outcomes of the combat.

[Ecthelion]

FYI, I just added a new version of the RM Combat Tester to the vault. It changes quite a lot of code, so that I hope I did not introduce any severe bugs  .