Withdraw Mod

This modification makes withdraw in combat more likely and stategically important.

http://forums.civfanatics.com/uploads/86216/Withdraw_Mod.zip

The planning thread is here:
http://forums.civfanatics.com/showthread.php?t=212617

in that thread, you can see the thought process. The basic idea is that a top-tech great-general unit can hit 95% withdraw chance.

Tactically, making flanking units to "soften up" enemies becomes very powerful. This boosts the imporance of anti-cavalry troops.

Ranged troops gain the ability to build up a high withdraw chance, but it requires more promotions than mounted units do. This is through adding withdraw chance to the drill and barrage upgrade paths.

Detailed changes:


UUs where modified along with the core unit.

Enjoy.

Tested using the 1000 AD scenerio with the mongols. Great General Kenshiks are nice, and swarms of flanking kenshiks work wonders at overrunning the chinese.

Playtesting:

I tried out a mix of this with insane barbarians (which gives lots of early combat). Sadly a lack of horses meant I wasn't able to test it out with chariots, but I did give archers a large workout.

The improved Drill may still not be good enough. Combat is better in many cases: the drill unit's victory+withdraw chances still doesn't dominate the combat units victory+withdraw chances in most cases.

Defending, in theory I think a high-drill unit takes less damage on average.

Let's compare Drill III in a high-defence hill city to City III.

A Drill III archer in a 60% culture city on a hill has 7.8 strength and 5 to 9 first strikes.

Damage: 21 for archer, 18 for swordsman.

Against a combat 1 swordsman it is 6.6 vs 7.5. 53% of first strikes hit.

An average of 7 first strikes means an average of 78.2 damage before the swordsman reaches the archer.

Post-first-strike, the archer does 1.33 damage for every damage the swordsman does, which means the archer takes an average of 16 damage killing the swordsman.

A city defence III archer vs a swordsman with combat I in a 60% defence city:
75+60+50+50-10 = +225 archer strength, or 9.75 strength, vs 6.6 strength.

Damage 24 for archer, and 16 for swordsman.

59% of the first stike hits, for an average of 14 damage.

Post-first-strike, the archer hits 1.47 times as often, for 50% more damage, or 2.22 times as much damage as the swordsman.

This means the city defence III archer takes an average of 38.8 damage defending against the swordsman.

...

So the drill promotion doesn't work that badly at efficiently soaking attacks.

What it doesn't do is give the same level of boost to an actual successful kill.

...

I'm thinking that with the new higher withdraw chances, seige engines should be weakened somehow (maybe weaken catapults and trechepults by 1 strength each, cannons by 2 and artillery by 3?), and maybe toss on an extra -10% city attack on all mounted units (bringing horse archers up to -20%).

Balance comments:
Drill is actually viable. A strength 3 archer unit with Drill IV has a 70% chance to beat a strength 4 defender (and only a 20% chance of dieing).

With combat 4, that archer would have 4.2 strength, just enough to hit a 60% to 65% victory chance.

Before these changes, Drill was a pretty gimpy promotion path. Now, especially for protective civs, it is the main combat promotion path of archers.

Bugs in above post:
Missed the +25% withdraw on longbows.
Gunpowder units don't get Drill in base Civ4/Warlords. Oops!

Planned Changes from Playtesting: (not posted)
An extra -10% against cities added to all mounted units.
Cut the strength of siege engines.
Cat: 2 str +50%/+125% Attack/Defense cities (up to 4.5 str)
Treb: 3 str +75%/+150% A/D cities (up to 7.5 str)
Cannon: 6 str +100%/+175% cities (up to 16.5 str)
Artillery: 9 str +125%/+200% cities (up to 27 str)

All siege now defend first against siege, which means you can attack the siege engines of the force that is tearing down your cities defences.

I might tone the city defense bonus of siege engines down and instead let them use terrain defences.

Collateral damage of siege engines is limited to 30% of the target's HP (ie, you can beat targets down to 70% health, but no further).

Collateral damage of other units is limited to 20% of target's HP.

Siege engines are no longer immune to collateral damage, except Machine Guns.

Added Drill to gunpowder units.

---

I'm tempted to do "archer duels" as well as siege duels, where ranged units fight each other before they fight melee units.

Ie, Archers defend against Archers and Gunpowder.
Gunpowder defends against Archers, Gunpowder, Armor.
Armor defends against Armor, Gunpowder.

This would allow you to build "anti-archer" archers, and attempt to gain archer superiority in siege. The defenders could attack your archers without having to fight the melee units.

Giving machine guns +50% against mounted. With modern high-rate-of-fire weapons, mounted troops become completely obsolete.

...

I'm also tempted to boost siege engine withdraw chances even further. With the ability of enemy siege engines to attack your siege engines, having damaged siege engines sucks badly. Plus, given the poor direct-combat abilities of siege engines in this mod, you will need support troops.

Tactical changes:
You need siege engines on both attack and defence. The combat abilities of a siege engine mainly let you survive enemy siege engine duels. A siege engine will lose against an even-tech defending unit in most cases, unless it is buffed to hell and gone with city raider.

A city-raider buffed siege engine makes a decent unit to take out cities. 9 strength +200% city attack artillery can defeat a 1 turn fortified city defence 2 strength 20 infantry more than half of the time.

The city raider 3 artillery is even with a defending artillery, but wins if the defending artillery has taken collateral damage and the attacker hasn't.

The use of cavalry or armor (or archers if you are desperate) to soften up targets with flanking attacks becomes important. You can no longer just siege a target down to half-health and then win: instead, you will want to collateral the target down, soften the target up with high withdraw-rate troops, then smash the target with your strong troops.

All mounted troops can hit over 50% withdraw rates with just two promotions (Chariot: 55%, Knight: 60%, Horse Archer: 65%, Cavalry: 70%, Chopper: 70% (+5% with a 3rd promotion)).

Longbowmen can hit 45% withdraw with 2 promotions (drill I and II), and are not that far beyond catapults.

Horsebackriding is worth going for, dispite it's high cost, because horse archers fill in a tactical niche in fighting wars. 65% withdraw chance!

I need to be more systematic about siege engines, so they aren't too strong or too weak.

The damage done by a "suicide charge" of a siege engine against a stack is:
100 * A/D * [(A*3+D)/(D*3+A)]^2 + 20*T * A*C/(A*C+D) * (A*C*3+D)/(D*3+A*C)
where A is attack, D is defence, C is the collateral multiplier, and T is the number of collateral targets.

When working out production tradeoffs of an assault, you get to divide by the survival rate: a 55% retreat rate (Cannon with BarrageII) gets to do 2.22 suicide-attacks on average before being destroyed.

Two things are included to make pure-cannon assaults less optimal: first, collateral damage only does 30% damage at most.

Second, making the siege units themselves weak, and vunerable to enemy siege unit duels.

I'll assume Barrage II, and City II defenders fortified in a city with it's defence removed.

Trechpult (3+75%) vs Longbow (6+95%) extra 4 targets
Direct: 3/7.2 * 100*((9+7.2)/(7.2*3+3))*((9+7.2)/(7.2*3+3))
18

Collateral: + 4 * 20*4.5/(4.5+7.2) * (13.5+7.2)/(21.6+4.5)
24.4

Total: 42.4.

Retreat: 50%, so 84.8 damage per dead trec.

Longbow costs 50.
Trech costs 80.

That doesn't seem out of line. So if you throw trech's against longbow defenders, you will end up spending more production on softening up the targets than the percent of damage you do to the targets. (Ie: to reduce 5 longbowmen to 70% health takes more production in lost catapults than 30%*5 = 1.5 longbowmen).

Next:
Catapult (2+50% @45% retreat) vs Archer (3+120%)
Cannon (6+100% @55% retreat) vs Rifle (14+70%)
Artillery (9+125% @60% retreat) vs Infantry (20+70%)

And I want to check "field battles" of catapults against chariots/horse archers/knights/etc.

Lastly, make sure that mounted vs city and mounted vs anti-mounted don't end up being overly optimal. I don't want to replace "cats ftw" with "mounted ftw" or "archers ftw".

The catapult phase should be inefficient, cost to damage wise, because the purpose of the phase is to "soften up" the enemy targets so your other units can get easy wins.

The mounted phase, against damaged targets, should also be inefficient.

Finally, the conquest phase, where you attack the barraged and flanked troops with units that are likely to win, should be efficient.