Evade settings being changed?

TheGrayMouser · Post by **TheGrayMouser** » Tue Sep 13, 2011 11:00 am

Fedem wrote:TGM by saying: "so, currently it appears that BOW cavalry will NOT evade from a BAD combat if the attacker is drilled! (at least with quality being the same etc as per my above tests"

You meant Bow cavalry will NOT evade from a BAD combat if the attacker (Heavy Armored Lancers) is UNDRILLED right? If drilled they get the hell out of there

Cheers.

Yes!

TheGrayMouser · Post by **TheGrayMouser** » Mon Sep 26, 2011 5:24 pm

***Just bumping this as i see Keith is back

ran a series of tests in the editor , ran multiple times 10 units vs 10 in three batches

*** all Bg's on both sides were superior*** test subjects all set to AI evade**all tests in nice clear hexes all around***

test one (lance-sword Knights HA vs Armoured BOW Sword cavalry) odds 59-18 % win/loss for the knights
result NEVER , not once did the bow cavalry evade

test two (lance-sword Knights HA vs Armoured Sword cavalry) odds 59-18 % win/loss for the knights
result: cavalry evaded EVERYTIME

test three: (lance-sword Knights HA vs Armoured Light Spear Sword cavalry) odds 59-18 % win/loss for the knights
result: cavalry evaded EVERTIME

Interstingly enough the cavalry units in all three situations would have been DOWN a POA in impact , and DOWN a poa in melee
Why then do the units that have bows in addition to swords think they can fight knights??? The combat odds are exactly the same!

Also thanks to Fedem whom uncovered that if the Knights in question are drilled, then the bow armed cavalry evade.

So, appears "AI evade stances" are a little wonlky it specific cicumstances after the lates patch....

This also effects custom/ historical scenarios as well as the AI doesnt appear to change the "stances" it has but depends on the AI Evade logic to do the , uhhh, logical thing. Which stinks because now you can charge bow cavalry with lances fully in the knowledge that they will stand ( and die)

keithmartinsmith · Post by **keithmartinsmith** » Mon Sep 26, 2011 6:19 pm

Will look at and review this. Thanks Keith

batesmotel · Post by **batesmotel** » Mon Sep 26, 2011 6:49 pm

This almost sounds like the evade decision is being made based on the relative cost of the BGs.

Chris

TheGrayMouser wrote:***Just bumping this as i see Keith is back

ran a series of tests in the editor , ran multiple times 10 units vs 10 in three batches

*** all Bg's on both sides were superior*** test subjects all set to AI evade**all tests in nice clear hexes all around***

test one (lance-sword Knights HA vs Armoured BOW Sword cavalry) odds 59-18 % win/loss for the knights
result NEVER , not once did the bow cavalry evade

test two (lance-sword Knights HA vs Armoured Sword cavalry) odds 59-18 % win/loss for the knights
result: cavalry evaded EVERYTIME

test three: (lance-sword Knights HA vs Armoured Light Spear Sword cavalry) odds 59-18 % win/loss for the knights
result: cavalry evaded EVERTIME

Interstingly enough the cavalry units in all three situations would have been DOWN a POA in impact , and DOWN a poa in melee
Why then do the units that have bows in addition to swords think they can fight knights??? The combat odds are exactly the same!

Also thanks to Fedem whom uncovered that if the Knights in question are drilled, then the bow armed cavalry evade.

So, appears "AI evade stances" are a little wonlky it specific cicumstances after the lates patch....

This also effects custom/ historical scenarios as well as the AI doesnt appear to change the "stances" it has but depends on the AI Evade logic to do the , uhhh, logical thing. Which stinks because now you can charge bow cavalry with lances fully in the knowledge that they will stand ( and die)

TheGrayMouser · Post by **TheGrayMouser** » Mon Sep 26, 2011 7:06 pm

It might, except that if you take away the bows from the evader they WILL evade ( and bows cost alot) I personally think ( ok 100% guess) is that something got mucked up when they added in the cavalry rear rank bow/ x-bow class weapons for knights and cavalry units.

batesmotel · Post by **batesmotel** » Mon Sep 26, 2011 8:30 pm

As a quick test for my relative points cost, do Ottoman Qapu Khalki evade from an elephant? Elite drilled bqow sword armoured cavalry is more points than an elephant (and might even win the melee a fair amount of the time with re-rolls.

Chris

TheGrayMouser · Post by **TheGrayMouser** » Mon Sep 26, 2011 8:57 pm

batesmotel wrote:As a quick test for my relative points cost, do Ottoman Qapu Khalki evade from an elephant? Elite drilled bqow sword armoured cavalry is more points than an elephant (and might even win the melee a fair amount of the time with re-rolls.

Chris

AP's should (hopefully) have nothing to do with evades! As per my post/test results I was going with the assumption that AI evade was based off the % chance of winning the impact combat so really the only things that matter are: Quality(rerolls) and Impact weapon(poa's) and of course unit state ( ie disurpted etc which wasnt part of the tests anyhow) .

Hmm, that test might not be a qualifier as elites with their rerolls might very well have closer to equal odds vs an elephant in impact , despite the pac being up a POA. Hmm of course the Qupu cavalry will be dispordered from pac stink so likly the odds will be similar o the other tests ie 59-18 in favour of the pacs so the Quupu Cavalry SHOULD evade!

In all my test both attacker /defender were same quality level .

Of course all this could mean nothing at all as the manual states impact and melee?? will determine evades..... This to me makes no sence and my tests dont support that either , once in melee a bow armoured sword cavalry is exactly the same as a spear sword armoured cavalry , which same as a knight -armoured- sword etc etc , at least in terms of rerolls POAs etc