Log in

View Full Version : Archery Study



Red Harvest
03-13-2005, 09:37
I did some extreme testing of archery tonight (63 missile attack vs. no armour, no shield peasants.) I wanted to get a feel for the "accuracy limit" of archers, distance effects, target formation effect, and whether or not the "lethality" stat played any part. I've seen some incredibly accurate archery at range and wondered what the limits were to this.

Conclusions first:
1. There is no apparent raw "accuracy" factor for archery. Accuracy seems to be determined by collision calcs. and attack/def without an accuracy fudge factor. I've seen up to 50% of a volley produce kills vs a moving target. 80 archers (+2 non firing officers), 40 kills in one volley. (Remember this is EXTREME testing.) When I refer to kill rate it will be based on the archer count, not the target count (target was intentionally more numerous than the archer unit.) Considering that many arrows would hit forward targets at the same time in a single volley, at least 50% of the four deep target should be behind a "meat shield."
2. There is no apparent factored distance reduction on penetration, lethality, or on accuracy. (There is a chance that the very high attack is influencing this, so it warrants another look. However, I've noticed the same basic appearance in campaign from the beginning.) Reduction in lethality at distance seems to correlate only with arrow trajectory and whether or not the arrow is in the invisible hit box. So there is some reduction in accuracy at distance for geometric reasons but it correlates with additional "miss area" available due to depth playing a factor rather than just frontage (1 dimension vs. 2) or a fudge factor. For my testing, point blank accuracy was ~46% while 150 meter accuracy was ~30%. If impact energy, dispersion, etc. were considered I would expect kills to fall several fold at max. distance.
3. Target formation spacing has less impact than expected. I doubled horizontal and vertical spacing, so that the target formation took ~4 times as much area (2x2.) This resulted in a reduction of kill rate at 150 meters from 30% to about 19%. This suggests that frontage and formation depth were more important than total spacing, even at max range.
4. The archer primary weapon "lethality" stat had no noticeable impact on achery kills in the range 0.01 to 1. That is a 100 fold factor and should have been enough to compensate for the crazy high missile attack.

Methodology:
Grassland map, midday, calm, 3 tests per setting.
Large unit size, medium difficulty
Achers: Roman Archer Auxilia modded to 150 meter range, 63 missile attack.
Targets: Greek peasants set to 1,1,2,2,4 square formation. No armour, no shields. For formation width/depth testing stats were 2,2,4,4,4 square formation. The target was moving in all tests.

P.S. I've got another thing to try...noticed it in the projectile stats and want to see how it works.

Kraxis
03-13-2005, 13:54
Try going back to Horde for the peasants... They might become less 'targetable'. ~:cool:

Anyway I'm surprised the lethality didn't do anything... But maybe it is like the MTW lethality for archers, how many HP it will drain. This will not explain the very low lethality but perhaps there is a minimum of 1 HP drain? Try upping to 4 lethality and shoot at Spartans or Bull Warriors.

hrvojej
03-13-2005, 14:46
Also, frequency of the attack (second to last number in stat_pri, usually 25) has no effect on time between volleys. But neither does the stat_fire_delay according to the tests I've done previously, which is slightly weird. If you figure out how to increase the time between volleys, don't hesitate to post about it.

Cheers,

Red Harvest
03-14-2005, 06:22
I just retested with missile attack of 15 (vs. 63 before). This ~4 fold decrease caused ~50% reduction in killing power at 150 meters, from 30% archer kill rate, to 16%. I attribute this not to any change in accuracy, but instead the reduced attack value.

Kraxis
03-14-2005, 14:14
Seemingly archery is harder to mod compared to MTW... And yet easier at the same time. We had quality modding in MTW but lacked the quantity (could only make three types of bows), here we have lots of quantity (individual modding) but lack the quality.

Uesugi Kenshin
03-14-2005, 15:44
Interesting test! I look forward to see what happens if you up the lethality and shoot at spartans. Good luck figuring it out.

Red Harvest
03-14-2005, 21:43
I tried 63 attack and 9 lethality vs. Spartans. First volley = 2 kills.

I then tried cutting lethality to 0.5. First volley = 4 kills.

Conclusion: The difference in kills was random variation against a 2 hit point unit (where two hits and kills must register on a single man in the first volley.) If the 9 lethality worked at all it would have killed at least a dozen spartans.

They were mown down like wheat in subsequent volleys since almost all front rankers had lost their 1st hit point.

Kraxis
03-14-2005, 22:03
That downed that theory...

Uesugi Kenshin
03-15-2005, 03:12
What about against a testudo? I know that randomly some guys seem to get hit, I bet this is based on chance, how well they are covered and so on. Would you mind testing that as well?

Red Harvest
03-15-2005, 03:50
What about against a testudo? I know that randomly some guys seem to get hit, I bet this is based on chance, how well they are covered and so on. Would you mind testing that as well?

I have no idea how to get the AI to form testudo... And if I let the AI be the archers it tends to charge my line...stupid AI. So vs. the AI I really can't perform a "standardized test."

Kraxis
03-15-2005, 11:24
Hmm... Yes that is a problem but I think you could persuade someone to help you in an MP test.

Kraxis
03-15-2005, 12:31
Did a little test myself, only to debunk or confirm the charging AI.

Had a Legionary First Cohort against a unit of archers from Pontus. They did indeed close up dramatically, but when inside the pilumrange (and just outside the chargethreshold) they began to shoot.

It appears that Testudo only grants the shieldbonus to the exposed top and sides, not any other more fictive bonus, like Open formation grants a point of armour. But then again I mostly saw the arrows hit the front or the men on the sides.

Productivity
03-15-2005, 12:52
I have no idea how to get the AI to form testudo... And if I let the AI be the archers it tends to charge my line...stupid AI. So vs. the AI I really can't perform a "standardized test."

If you want I can be your opponent to fire missiles at, send me a message using any instant messenger and we can (try) to work out a time.