archery tests (are Samurai archers worth it? etc)

Hypothesis: I wanted to see if Samurai archers were worth it compared to Ashigaru ones. The null hypothesis was that there was no difference in their missile kill rates. It has been suggested that Samurai archer fire is "armor piercing", unlike Ashigaru fire, so I wanted to test for differences against targets of different armor.

Method: I tested one unit of archers under my control against one melee unit under AI control. I used the Crossroads map, which is fairly flat - deploying my archers back near some trees and allowing the AI to fire at will. The AI melee infantry started running when it got into range. I paused as the melee infantry hit the archer line to record how many had survived and how many had died to archery. I repeated each trial four times, comparing bow ashigaru and bow samurai against, variously, yari ashigaru, katana samurai and naginata.

Results: Ranged kills in each trial (e.g. 34 Yari Ashigaru died to missile fire in the first time I had them attack Ashigaru archers).


Discussion: There was considerable variation in the results - sometimes this was due to the contest involving one more or less whole volley. This may have been partly due to minor differences in the set up of the tests (the location and alignment of the archers; possibly even the weather). The number of trials is too small for statistical testing but the results are interesting.
The Samurai archers had 22.6% higher missile kill rates than Ashigaru archers against Yari Ashigaru, but 39.3% higher against Katana Samurai and 58.9% higher against Naginata.
Conversely, the Katana Samurai (armour=5) suffered 34.7% fewer casualties than Yari Ashigaru (armor=2) against Ashigaru archers and 25.8% fewer against Samurai. The Naginta (armour=9) suffered 54.1% fewer casualties than Katana Samurai against Ashigaru archers and 47.6% fewer against Samurai.

Inferences:
Despite the smaller unit size (90 vs 120), the Samurai archers kill more than Ashigaru archers. The difference is modest when fighting virtually unarmoured Yari Ashigaru, but substantial against armoured Samurai and even more, the heavily armored Naginata. Subjectively, I would conclude that Samurai archers are "worth it" in your main armies by the mid to late game when most of your enemies are samurai. Added to this, Samurai are much more melee capable (they always beat the Yari ashigaru whereas the Bow Ashigaru always lost).

The comparison of targets by armor suggests that armor upgrades (e.g. from smith provinces) are valuable. The +3 from an armorsmith could cut the ranged casualties of your Yari ashigaru by a third.

Finally, naginata rock - taking about half the casualties of katana samurai to missile fire.

Possible further research
I tried using fire arrows (Ashigaru vs Ashigaru) but observed no discernible difference in kill rates. It may be their main (only?) effect is to lower morale (and burn structures). Usage appeared erratic though - sometimes only one full volley of fire arrows went off, another time, three volleys. It may be that it is better to press fire arrows when the enemy target is in mid-range to maximise time on target (I pressed when they were just out of range).

It would be interesting to see the effect of experience upgrades.

I wondered if manually targeting arrows might give better results than auto-fire. Typically, at maximum range, only one or two archers appeared to fire. The first big volley comes quite a while after the target has passed the red maximum range line. I wonder if this is an unavoidable delay or if autofire has additional lag.

I've been wondering this since day one of release.

I agree that much more testing is needed, but at least, initially, it appears samurai archers are best suited for an anti-armor role.

Good work econ! :yes:

Out of curiosity, what about monk archers? Was there a reason you didn't include them or is that just the subject of another discussion?

I've also noticed that only a couple archers in a group will fire at long range. My interpretation was that the archers in range fired first (in the front rows) but in fairness, I'd only noticed it in passing and hadn't paid a lot of attention.

These are Samurai Bow archers we are talking about,they can be far useful if used properly,the bow ashaguri are excllent archers(They make a good killing)Katana Samurai are a bit useless,the Cavalry is of the Takeda is great.

From the 'pedia, missile stats are listed:
Ashis: Accuracy 25, Reload 20, Range 150
Samurai: Accuracy 40, Reload 60, Range 150
Monks: Accuracy 50, Reload 65, Range 175

Do we know how the Reload stat relates to time between shots? The Samurai reload is considerably greater than the Ashis. Can you run the same test, but base it on number of volleys (pause after X volleys have been fired)? That would give you a better indication based on total shots fired (volleys x men in unit).

I'm also a bit curious about how upgrades change the equation. In a fully developed crafts province, you can easily get 60+ accuracy bow ashis.

I have a feeling they may well be more cost effective once upgrades are factored in.

There's no doubt about the cost effectiveness of bow ashigaru - even without upgrades, they cost half the upkeep and are only 28% less effective than bow samurai against katana samurai. But for your main army, you may want to pay extra to get the best you can.

Right now, I'm thinking of going for a mix of bow ashigaru and bow samurai in my mid to late game frontline armies. Say 3 ashis to skirmish in loose formation and take the casualties; and 3 bow samurai behind the lines to add weight to missile duels. Monk archers might well be even better in the latter role (because of the extra range).

To investigate many stats like accuracy upgrades, you'd have to start modding the unit stats. I would not be averse to doing that (e.g. I'd be very curious to compare the smith upgrades for melee, armor vs attack skill), but don't know how at the moment. If anyone with modding skill can explain how or point me in the direction of a "how to..." thread, I'd be grateful.

Econ, very cool work... it must take a lot of painstaking attention. Everyone here thanks you for the time spent on it!

One factor not being mentioned is, the quality of troops in a given space. Your 90 bow samurai only take 75% of the space of bow ashigaru on castle walls, or in your formation on the open field. IF all things were equal, this means they're 33% more effective, at least in one sense of the word. Then we can add Econ's findings on top of that, and they really start to look good. After all, why would the designers make smaller units (in terms of people) cost more?

As a bit of an aside, and contrary to what I said above, my impression is that bow warrior monks are not worth it (bow samurai are the happy medium), but naginata warrior monks ARE worth it, relative to spear samurai or ashigaru.

Let me add something.

You probably did only "standing" - tests. There is a curve with is very flat in the early stage concerning accuracy. Lets leave reload time out.
Its one thing to look at it, if units are standing, its completely different, if units running/walking.

From my experience and tests, there is a soft-cap at around 70-80 accuracy. At this point the accuracy starts to have some feel-able effect on the velocity.
This means, that arrows start to turn into cruise missles and move to hit the target, while archer with low accuracy will shot behind (when the units moving towards you).

You can see that in one of my posted vids (wolves vs aggony II) there is a moment in the game, where i document how my general dodge a full volley of the enemy archer.


Theory is good, from a practical POV, you have much better results, if you use monk archer with high accuracy, they really can "stop" units running to attack you.
You can spike this with inspire from you general.

As direct test, try ashi archer and monk, you can also zoom in and see, how arrows start flying behind the unit.

As far as only having a few archers, at first, fire when an enemy unit enters the firing arc, I believe that is a case of 'finding the range', if you will.

My question is, why do archers have to reposition themselves to fire at another target that is well within their arc. It seems to me that archers (Ashi and Samuari) will only fire if a unit is DIRECTLY in front of them, thus the almost constant repositioning and alot of time wasted that would be better used to shoot an enemy unit that is not already engaged in melee with your own troops. I mean, I dont see why they refuse to fire at a target that is 10 degrees to the left or right.

While I'm kind of ranting I'd also like to point out that archers seem to be almost unable to follow orders, or are....whats the word Im looking for.....retarded, thats it. When I micromanage archers I most of the time will have them with 'Fire at Will' off and tell them to stop shooting at a unit thats engaged in melee in order to minimize friendly fire, and they will stop for a second, probably reposition (again) and go right back to happily pounding on the very same unit I told them to stop firing at. Also Id love to have a 'Fire at this area' command when Im defending bridges.

If only there was the Roman practice of Decimation. I know who I'd pick out first.

Agreed, Phog. Thanks for the insight on why they move around for no reason.

Another favorite that I'm sure you've seen: Your 10 bow samurai units (900 archers) all fire at the 1 remaining bow ashigaru individual if you put them on free fire, even if impossible to hit him because behind something, when there are lots of other targets. Your only choice is to select one archer unit to do away with that last individual.

If you order your 10 units to shoot a selected target (with autotarget off), maybe you're lucky if half your units target what you selected.

Anyway. The consolation prize is that, humans always still rule. :grin3:

I don't mind TWS2 being lame in "open warfare"... it's one of the few games that even tries. Humans can run circles around a million lines of code.

I look forward to better AI.... but am a little scared of SkyNet.

RK

I'd really be happy to see if Samurai arrows actually are "armor piercing".

That would require a test where an ashigaru archer has the same accuracy and reload as the samurai unit, which I suppose is possible to do via valor and accuracy bonus?

There must be a way to mod it.

But I'm still just enjoying learning and playing TWS2.

Quote:

---

Originally Posted by Rothe

I'd really be happy to see if Samurai arrows actually are "armor piercing".

That would require a test where an ashigaru archer has the same accuracy and reload as the samurai unit, which I suppose is possible to do via valor and accuracy bonus?

---

Maybe I am wrong, but it seems to me that my tests are sufficient. If Samurai and Ashigaru had the same chance to penetrate armor, then their relative kill rates should be the same for targets of any armor class. The fact that the Samurai have higher kill rates relative to Ashigaru against armored opponents implies they must have some additional efficacy against armor - i.e. be armor piercing.

Let's say the expected number of hits on a target in a given time period is Hs for Samurai and Ha for Ashigaru. This number would depend on accuracy and reload but should be same whatever the armor class of the target.

The chance for a hit to kill (i.e. to penetrate armor) is Ks for Samurai and Ka for Ashigaru. If Samurai are not armor piercing, this should be the same for Samurai and Ashigaru (Ks=Ka).

The number of kills is Hs.Ks and Ha.Ka for Samurai and Ashigaru respectively.

So for a lower armour class target (1) and a higher armor target (2), the ratio of the number of Samurai kills to Ashigaru would be Hs.Ks(1)/Ha.Ka(1) and Hs.Ks(2)/Ha.Ka(2). If both are equally (in)capable of piercing armor, we have a ratio of Hs/Ha for both types of targets.

Even if I am wrong, I am not sure why you would want to know more about whether Samurai arrows are "armor piercing"? The tests show their superiority over Ashigaru is greater, the more armor the target has. Whether that is because they can pierce armor or because Ashigarus' aim starts to shake when faced with armored opponents, the implications for your recruitment decisions are the same.

Quote:

---

Originally Posted by econ21

Maybe I am wrong, but it seems to me that my tests are sufficient. If Samurai and Ashigaru had the same chance to penetrate armor, then their relative kill rates should be the same for targets of any armor class. The fact that the Samurai have higher kill rates relative to Ashigaru against armored opponents implies they must have some additional efficacy against armor - i.e. be armor piercing.

Let's say the expected number of hits on a target in a given time period is Hs for Samurai and Ha for Ashigaru. This number would depend on accuracy and reload but should be same whatever the armor class of the target.

The chance for a hit to kill (i.e. to penetrate armor) is Ks for Samurai and Ka for Ashigaru. If Samurai are not armor piercing, this should be the same for Samurai and Ashigaru (Ks=Ka).

The number of kills is Hs.Ks and Ha.Ka for Samurai and Ashigaru respectively.

So for a lower armour class target (1) and a higher armor target (2), the ratio of the number of Samurai kills to Ashigaru would be Hs.Ks(1)/Ha.Ka(1) and Hs.Ks(2)/Ha.Ka(2). If both are equally (in)capable of piercing armor, we have a ratio of Hs/Ha for both types of targets.

Even if I am wrong, I am not sure why you would want to know more about whether Samurai arrows are "armor piercing"? The tests show their superiority over Ashigaru is greater, the more armor the target has. Whether that is because they can pierce armor or because Ashigarus' aim starts to shake when faced with armored opponents, the implications for your recruitment decisions are the same.

---

That is partially true, but for multiplayer it is interesting to see how much more effective potential armour-piercing arrows are against higher armoured targets such as Naginata Samurai who have a whopping 9 armour and are very very hard to kill with regular arrows regardless of accuracy.

You guys might be forgetting one crucial fact: An improved accuracy score may simply decrease the effectiveness of armour (if you'll allow me to geek out for an example: Legolas telling the elf archers at Helm's Deep where the armour on the Uruk-hai is weakest.) The real question I have to wonder: Are bow hero units worth it? Also, I don't think you should measure the kills per volley, but perhaps the kills in a given 30 seconds period or something, to allow for reload speed to come into account.

Quote:

---

Originally Posted by Rocinante

That is partially true, but for multiplayer it is interesting to see how much more effective potential armour-piercing arrows are against higher armoured targets such as Naginata Samurai who have a whopping 9 armour and are very very hard to kill with regular arrows regardless of accuracy.

---

Well, if you look at the test results, you see that difference between kills by samurai archers and kills by the ashigaru archers was greater the higher the target armor...
So, whatever the reason, the samurai archers perform better against armor and the highger the target armor, the bigger the gap between samurai and ashigaru performance

Quote:

---

Originally Posted by RedKnight

One factor not being mentioned is, the quality of troops in a given space. Your 90 bow samurai only take 75% of the space of bow ashigaru on castle walls, or in your formation on the open field. IF all things were equal, this means they're 33% more effective, at least in one sense of the word.

---

Very true. However, by the same token, casualties taken 'hurt' the effectiveness of the unit too, with each casualty for Ashigaru reducing total unit effectiveness by 1/120th while each casualty for Samurai reduces it by 1/90th.

Quote:

---

Originally Posted by Daevyll

Very true. However, by the same token, casualties taken 'hurt' the effectiveness of the unit too, with each casualty for Ashigaru reducing total unit effectiveness by 1/120th while each casualty for Samurai reduces it by 1/90th.

---

But killing Ashigaru is so much easier than killing samurai.
4 armour instead of 1, and a higher melee defense also.

my opinion ,ashigaru archers,then bow samurai,then monk archers,fire faster ,fire more means kill more!!!