Skirmisher cav not very good at skirmishing?

[aaronp]

I don't have much experience using skirmisher cav in M2, is this an engine issue? I noticed that the missile stats for these units are reasonably high, but the javelins very rarely cause casualties. Mostly I use them to tie up the enemy cav in melee.

Recently one of my client kings in North Africa was attacked and I wanted to confirm. I had several units of local skirmisher cav, several units of foot archers, a unit of slingers and a couple of units of leftover infantry from Italy. The enemy was almost all skirmisher cav and I decided not to engage in melee. The enemy's skirmisher cav stupidly bunched up into a ball, making them a good target for missiles. My archer units got from 90-125 kills, and my skirmisher cav units got around 20-30. This was a long battle that ran about 3/4 of the way through my time. In a "normal" battle where the skirmisher cav are supporting infantry, they would get next to no kills with their javelins.

[QuintusSertorius]

Where are you attacking from? Javelins are most effective from behind, or at least in the unshielded flank.

[aaronp]

Hmm, that's a good thing to note that I didn't consider. My cav units were in a semi-circle around theirs (not bunched into a single mass like theirs) but I do think they were focused more on my cav because they were closer, so the archers would definitely have had an advantage there. I dunno, maybe my perception is wrong and I should stop using them so much for melee

[Septentrionalis]

Where are you attacking from? Javelins are most effective from behind, or at least in the unshielded flank.

It makes a difference which flank I attack from?! I never thought that the tactical modelling is that detailed. Thank you for that information, QS.

[BailianSteel]

M2TW is old, but old games can have a surprising amount of detail. Units have a shield value that only protects them from the front. There are also height and terrain advantages, I once won a battle I thought unwinnable because my enemy had to scale a mountain to attack me.

That being said, it does seem like skirmisher cav is mostly there to pad out numbers of horse, provide opportunities for flanking attacks on the cheap, and hopefully keep the powerful enemy units distracted.

[Hellenikon]

There is no unshielded flank in M2TW, that was only back in RTW.

In M2TW:K, both flanks offer 50% of the shield protection. You can test it out easily.

Source:
https://www.twcenter.net/forums/show...1#post11406767
https://www.twcenter.net/forums/showthread.php?536361-Shields&p=11417460&viewfull=1#post11417460

This could be a very good thing to ask to the memory editing project, as it would very likely involve tweaking a small value in the hardcoded files of the game. If this was feasible to do, then we could restore how shields worked in RTW to the M2TW engine.

[QuintusSertorius]

I thought it was the same in M2TW; that's a weird thing for CA to opt to downgrade between engines.

[Septentrionalis]

There is no unshielded flank in M2TW, that was only back in RTW.

[...]

This could be a very good thing to ask to the memory editing project, as it would very likely involve tweaking a small value in the hardcoded files of the game. If this was feasible to do, then we could restore how shields worked in RTW to the M2TW engine.

Thank you for the information. Are you going to ask?

I thought it was the same in M2TW; that's a weird thing for CA to opt to downgrade between engines.

The programming stuff that I work with does not relate to game development in any meaningful way, so it is difficult to understand albeit fascinating. Could it be oversight and reverting back to an earlier code inadvertently? It is hard to imagine that there would be a significant computational cost. Of course, one would have to take into account whether the unit has a shield to begin with.

[k/t]

I can't believe people still believe that "unshielded flank" stuff in 2020. Especially someone like QS...

[Camcolit]

The first thing I do when coming back for a game of EB2 is edit the edu and increase the attack values of all light javelin units. Without it they are useless from the front, which is how they were historically used.

[moisesjns]

The first thing I do when coming back for a game of EB2 is edit the edu and increase the attack values of all light javelin units. Without it they are useless from the front, which is how they were historically used.

Also increase their Range so they actually don't get in a fight right away. YES it may look weird to see them throw and hit targets from unrealistic ranges such as 100 ft + but its not like the engine can represent their true usage of Attacking Foragers/ and hiding in trees/ forest.

[Camcolit]

Javelin changes edu attached if you want to use it.

All thrown weapon attack values increased by a flat 2 points except pilum which was unchanged (I think 20 is enough, soliferrum was only 22, now 24), and any javelin units that had 4 attack (lol Britain and Ireland backwaters). They are now 7. I don't play the region much so if you find this unbalanced you can reduce it but 4 attack seems laughable to me even against unarmored targets.

All thrown weapon range increased by 10 across the board. The original values tended to range from about 55 to 70 so it's an increase of 15% or so, bit more for heavier javelins, bit less for lighter. Nothing outrageous, the shortest range archers in the game still outrange the longest range javelins, but maybe you can get a volley off now before melee, or actually pull off a skirmish with skirmishers.

[TwigV]

Javelin changes edu attached if you want to use it.

All thrown weapon attack values increased by a flat 2 points except pilum which was unchanged (I think 20 is enough, soliferrum was only 22, now 24), and any javelin units that had 4 attack (lol Britain and Ireland backwaters). They are now 7. I don't play the region much so if you find this unbalanced you can reduce it but 4 attack seems laughable to me even against unarmored targets.

All thrown weapon range increased by 10 across the board. The original values tended to range from about 55 to 70 so it's an increase of 15% or so, bit more for heavier javelins, bit less for lighter. Nothing outrageous, the shortest range archers in the game still outrange the longest range javelins, but maybe you can get a volley off now before melee, or actually pull off a skirmish with skirmishers.

Whoa that is a laborious way of attempting to fix javelins (damage wise, yes range is fixable here). I've had the same thoughts come through my head - I even started to do calculations checked by maths to understand effects of facing and so on and be able to predict exact number of kills from volleys of javelins. At the end of the day the damage value(Random integer between 0 and listed value) is only there to overcome Armour(Random integer between 0 and listed value)+Shield(Random integer between 0 and listed value).

In the end, for all but the most anemic javelins (poor 7 attack javelins, maybe 12 attack javelins), the number you kill is actually dictated by the weapon legality, which is NOT in the units file. Go and mod the "descr_projectile.txt" file; go to the 'javelins and throwing spears' section and adjust the 'accuracy_vs_units' parameter. Looking at this number in isolation is a bit dumb because - there is a lot going on here - but that parameter basically sets the kill ceiling per volley. You can make projectiles do infinite damage in the units file, and you'll only kill about 15 in a volley from 160 units, but changing that parameter, will make the weapon kill beyond that.

I'm not sure why the devs took this design philosophy. In-game, reading the damage value on a javelin actually does not tell you enough. And looking at that value for different weapons will make you ask more questions like: why does the "rock" have a better accuracy (Read that as kill rate), than "glandes" which are lead bullets.......... *shrug*

[QuintusSertorius]

Uh, projectile lethality is inverse. Lower numbers are better.

[blackbirdgriffin]

I always find them quite usefull in small groups and I consider them as essentially light cav, I play with them by sending them ahead to harass CAV of screen for my other cav and once they're out of javelin they are great at disrupting enemy formation, tying down enemy CAV.

As long as they aren't bogged down in a fight and avoid charges from enemy cav they are good. They're biggest flaw is that they have low armor and shield and overall defense meaning they take loses in "big battles".

The only one I never use are the greek super light Hippokonkistai (not sure about the name) because they evolve in a highly armored envirronement and I mostly recruit them when I really need to scrap the bottom of the barrel in terms of recruitement.

But their Celtic counter part on the other side I really love using especially as a raiding force with my the soon to be Kingetos of my tribe or along a galatian recruited force when I play Pontos.

But then Again that's just me I'm sure some of the folks here have a few better ways of using them which I would be very glad to read and learn about.

[Camcolit]

Whoa that is a laborious way of attempting to fix javelins

What can I say, I like data entry

I am playing a KH campaign now and trying to use more skirmishers and also paying attention to the javelin armed infantry. I like how they are performing with the new values. Don't really want to mess around with projectile lethality. They are definitely getting more kills and are more comfortable to use with the longer range. Suits my taste but your mileage may vary.

[TwigV]

Uh, projectile lethality is inverse. Lower numbers are better.

Right, sure. I don't remember well, I looked at all of this a few months ago. I think at the end of the day I was content and stopped looking into it - I felt the balance was fine. I was also disappointed to learn from this thread that there was no shield facing mechanic. I think the main point I wanted to pass across to people modding their files was to look at the projectiles file to adjust things better. Also as I was trying to get the correct for for damage equations I found that there is a lot of misinformation on the boards about calculating how many are killed - the lethality number gets conflated with a few other variables and its not super clear. I didn't bookmark the useful posts sadly. The best way to do it is to write up the calculation in a maths package and check in game against expected results... maybe I saved something I'll check my laptop.

[TwigV]

Right, I redid some maths, and now I remember the subtleties. The way its designed right now: the damage value on javelins is pretty much linear with the kills it gets (I use Hoplitai as my pin cushin) for all but the worst armored units [in the case of equal number of javelins hitting defenders]. The fact that the highest damaging javelin is the soliferrum (dealing 22 damage + 4 a hidden 4 in the projectiles file), closly followed by the pilum (only 2 damage less) happens to be nearing the damage cap imposed by the accuracy means that for for those two javelins, you will expect basically the kills when raining on weakly armored units. The actual calculation (that I think is correct) resembles the following:

ExpectedKills = SoldiersAtt *ActualAccuracy (1 - AccuracyVsUnits)*
  Probability[Att >= (Arm + Shi) , {discrete uniform distributions}]

Where SoldiersAtt is the number of attacking soldiers; ActualAccuracy is the base missile accuracy provided in "battle_config.xml"; AccuracyVsUnits is accuracy_vs_units "descr_projectile.txt"; Attack, Shield and Armor values are taken from "export_descr_unit.txt"

The overwhelmingly dull thing to observe here is that misslie accuracy has a base 0.175 and that applies accross the board (jav, sling, bow), and the differences that are meant to be imposed by the accuracy_vs_units is largely lost. The assertion that the kill potential is inversely proportional to accuracy_vs_units is not correct, and I tested this by buffing glandes (setting accuracy_vs_units to near zero) and using my favorite Sphendonetai vs some Hoplitai to witness the expected 5 kills per volley (recall that for non-javelins the shield value is double from the front).

[TwigV]

I apologise for triple posting (I still haven't posted enough to edit my posts apparently, can't even go back and fix typos). I did some more testing by fiddling with the battle_config.xml and descr_projectile.txt to double check if my calculation is right.

This kind of thing really grinds my gears. Its clear to me that someone went to some effort to write sensible relative accuracies for all the ranged weapon types in descr_projectile.txt, and someone much later down the line decided that all kill rates were too high, and modified battle_config.xml to reduce ranged kill rates to 0.175, and melee kill rates to 0.35. For reference These numbers are 1 and 2 in mods like SS.

The end result is that the accuracy_vs_units in the projectile file are now essentially meaningless and a level of abstraction is lost - making all missile weapons anemic is a similar way across the board with the only way to balance it is by changing the crude damage value in EDU, since the its too hard to differentiate missile units. I'm going to take an even longer stab in the dark and guess this is why my dear slingers lost their amour piercing property in previous EB2 version, and are now solely balanced by a higher attack stat and and the old fire delay compared archers.

[TwigV]

..Still can't edit.

So what I said in the post above is not correct - I must have been in rain at some point when testing. AccuracyVsUnits really does do what it should. The expected kills equation approximates the combat with variables as noted quite well (with notable extra emphasis on the AccuracyVsUnits - I'd like to think the real equation exists somewhere - but there's modifiers for terrain;weather; distance that I'm not going into detail... although I'm finding distance doesn't matter as much as I'd expect it to.

ExpectedKills = SoldiersAtt *ActualAccuracy *(1 - AccuracyVsUnits)^10 *   Probability[Att >= (Arm + Shi) , {discrete uniform distributions}]