New Beta Version - November 9th (11-9)

[tu_79]

Well, I have been testing this in game and the flanking does not seem to work like the "New Flanking" that is supposed to be, but it seems to be working like the old flanking, though the values are lower, making flanking irrelevant..... By the way, the "Old Way" is the way it should be...

View attachment 508783

Above: There is one flanker supporting the attacker and no units flanking the attacker, so there is a +5% Flanking Bonus for the attacker.

View attachment 508784

Above: There is no flanker supporting the attacker and one unit flanking the attacker, so the defender gets the 5% flanking bonus.

View attachment 508785

Above: There is one flanker supporting the attacker and one defender flanking the attacker, so the flanking bonuses cancel each other out.

View attachment 508786

Above: There is no flanker supporting the attacker and no defender flanking the attacker so neither side gets any flanking bonus.

This all seems the way it should be, though not the way advertised in previous posts. My only complaint is that the flanking bonus now seems too small to really matter.

There's a case missing :
Use one warrior to attack two enemy warriors that are adjacent each other, but one enemy warrior is not adjacent to your unit. If the bonus is based around defender adjacencies then the unit that is behind the action might be counting against you.

 

[kkkitten]

Something else that's happening — probably linked to the walls/cannon issue — is that ships are wiping out cities with high defense numbers much faster than before. Has anyone else noticed this? Example: A green, 73-strength defended enemy city. 3 slighlt injured Bombardment 3 cruisers hit it once, followed by a vanguard ironclad, all backed by an admiral. They came within a hair of taking it the first turn... and I didn't have the city surrounded. If it had been, or one more shot by any sort of ship, and it would have fallen in one turn.

I had a game where I built a fleet of SoTL upgraded English cruisers that one-shot almost every enemy city in sight. Each barrage was ~1k damage to cities. I guess no trade agreements were made that day I'm gonna use the fixes posted here for now until G's back with his spicy hotfixes

Quick Update: I implemented the fixes posted here and reloaded the save where I was one-shotting cities aaand... my cruisers are doing chip damage (~30) now like they should, and my cities aren't 1-hitting fusiliers! Loving my drawn-out battles now

 

[Questdog]

There's a case missing :
Use one warrior to attack two enemy warriors that are adjacent each other, but one enemy warrior is not adjacent to your unit. If the bonus is based around defender adjacencies then the unit that is behind the action might be counting against you.

That is the last picture I posted. An attacker hitting a defender who has two adjacent units behind it and not next to the attacker. There was no flanking bonus applied to the defender or the attacker.

In all of the pictures, the defender has two friendlies adjacent to him, but he only got a flanking bonus when one of those units was next to the attacker and only got a bonus for that unit, not for the other adjacent unit that was not in contact with the attacker.

 

[Owlbebach]

easy. the old code was plain broken. didn't work and wasn't aligned with what the combat preview was showing.

now the new system: count the number of friendly units around the defender. subtract the number of enemy units around the attacker. multiply by your flankimg bonus (base + promotions). if positive the attacker gets the bonus. if negative the defender gets the bonus. like before this applies to melee combat only.

If new Flanking works the way ilteroi described - i am heavily in favor of this. I am 100% sure that old flanking code was broken since forever, i was saying this like 2 years ago, but nobody listened to me and payed attention to it. Maybe it has some bugs right now (though i didn't notice any in my current game), but this can be fixed. The idea described above is absolutely right.

I don't see rear units supporting a flank as AI friendly, or player friendly for that matter, because you'll find yourself moving a lot more units to adjacent before the attack. Normally either your units can be a next to an enemy, or they can't. Now I have to think about what unit to support with a unit who is ranged or can't attack anything. While you should probably do it anyway, now you to have move your entire army before making an attack just to maximize that flanking bonus. Or alternatively for example, say you have a 'straight' line of units, you can use 2 horseman who are behind that line and chain 5 flanking attacks by moving those 2 units 1 movement at a time between each attacker at a time.

I was doing it since forever. And it is exactly the way you are supposed to fight. If you didn't do it before - you didn't use you army to full power. Small things like this make difference. In fact i frequently move cavalry units back and forth just for the sake of increasing flanking bonus for another unit.

Which is another thing I don't agree with. I don't think I have ever heard a single person on this board complain that flanking was too strong, this is a pretty big change to just come out with no warning or discussion.
Further, this doesn't just nerf a combat mechanic, it fundamentally shifts the balance between melee and ranged back towards ranged.

I was that type of person. I was constantly saying that Drill completely suck compared to Shock. Shock3 units were at least 50% stronger than Drill3 units IF you use them right. (but apparently people refuse to learn to use them right)

 

[BiteInTheMark]

What's the point of the discussion? Questog showed the results of flanking in 4 different situations, and I didn't see any issue in this calculation. The defender didn't get any bonuses for adjectend units of the defender which didn't are adjectend to the attacker.
The only issue is the value of 5%, that's rather small, and makes flanking movements irrelevant, in comparison to all the other modifiers which are in the game. Give it a value of 10% for every delta, and then simply try it.

 

[YukiN]

What's the current status of the spearman line against ranged mounted units?

 

[kitfox]

UPDATE Units SET IgnoreBuildingDefense = '0'.

Which file is this in? Or do I have to append this to one of the mod files?

 

[tu_79]

Well, well. It looks like we were having reading problems.

As Questdog shows, current flanking works as we say it should work. But reading ilteroi post, one could think that it worked the opposite way. (Whom is a friendly unit? Friendly to the attacker or friendly to the defender?)

Now, having flat bonuses for flanking shows up how weak this promotion actually is. Flanking is rewarding in itself, both drill and shock units get the +10% CS for every flanking unit.
Having 3 shock promotions raised the flanking bonus a 75%.
1 flanking units became +17.5%CS (7.5% from promotions)
2 flanking units became +35%CS (15% from promotions)
3 flanking units became +52.5% CS (22.5% from promotions)
Now, having 3 shock promotions give a flat +15%CS (per flanking unit?), so
1 flanking unit gets +25%CS (15% from promotions)
2 flanking units get +35%CS (15% from promotions)
3 flanking units get +45%CS (15% from promotions)

I like conceptually the idea of a flat promotion. This makes units that especialize in flanking be good at flanking even with a small advantage. But maybe it needs to be a little higher.

 

[Padre19]

Which file is this in? Or do I have to append this to one of the mod files?

And put that at the end of \Community Balance Overhaul\Balance Changes\Units\UnitCombat.sql to catch all units with this unintended property.

 

[Gazebo]

Flanking is working correctly. You are all over thinking it. The two changes were:

- we added the attacker/defender delta, so you don’t ‘flank’ enemies if they could also ‘flank’ you if they were to attack you. The delta is new and works fine.

- we changed how the value is added to the combat bonus for the unit. It used to be a weird modifier, with even weirder promotion elements - now it works like all other combat modifiers.

That’s it.

G

 

[tu_79]

Little questions.

1. Is the combat algorithm able to move several units, fire, then move them back within the turn? I mean, things like moving a pathfinder to provide flanking to a melee unit, let the melee unit attack, then retreat the pathfinder.
2. Is the combat algorithm able to reuse within the same turn a unit that were set to attack a threat that has already been dealt with? Like, a spearman that was set to attack an archer, but this archer has been killed by another unit, so the spearman is free to attack another unit in the same turn?
3. Is the combat algorith able to move great generals/admirals before the shooting phase begins, and move it again to another hot point after the combat in the former point is over?
4. Is there a combat stance for crippling units rather than going for the kill? In some cases (comparing my attack/defense power in the zone vs the attack/defense power of the enemy) I may prefer to cripple enemy units so they can't attack at full power and are forced to retreat, so I don't have to face fresh units. (I know many players would think that the AI is not working properly if it didn't go for the kill, but sometimes it is just better to not kill first).
5. How does the combat algoritm deal with incoming pillagers (fast moving units, generally alone, that attack from the rear hoping to pillage as much as posible)?

 

[nekokon]

Ouch, I misinterpreted ilteroi's explanation Must have made mistake when counting units in my last game too

 

[Stalker0]

Flanking is working correctly. You are all over thinking it. The two changes were:

- we added the attacker/defender delta, so you don’t ‘flank’ enemies if they could also ‘flank’ you if they were to attack you. The delta is new and works fine.

- we changed how the value is added to the combat bonus for the unit. It used to be a weird modifier, with even weirder promotion elements - now it works like all other combat modifiers.

That’s it.

G

G, with respect, the reason that we are all "overthinking it" is because we have been given very little real understanding of the new mechanic. You just throw some change notes on us and left us to fend for ourselves

1) The change notes say "the bonus itself was reduced to 5%x the delta of friendly v. unfriendly units adjacent to the unit being attacked". This implies that flanking is focused on the defending unit now, which would be a change. But as you just said that is not the case, it is still based on the attacking unit it would appear.

I believe you are saying it is still based on the attacking unit, but instead of "attacker gets all of its flanking bonuses and defender gets all of its flanking bonuses" its "only 1 side gets a bonus based on a delta".

2) I think we all understand that the flanking modifiers like shock are different now (static additions as opposed to multipliers). However, your statement above suggests that the core 15% flanking modifier did not work "normally". Is this correct...in the original mechanic, was the 15% flanking bonus not applied like a standard combat modifier?

Aka if I had a Drill I unit with 1 flank in the old system it "said" I had a +25% bonus applied to my units' strength (10 drill + 15 flank). Was this not how the math actually worked?

 

[TwoPodRay]

G, with respect, the reason that we are all "overthinking it" is because we have been given very little real understanding of the new mechanic. You just throw some change notes on us and left us to fend for ourselves

1) The change notes say "the bonus itself was reduced to 5%x the delta of friendly v. unfriendly units adjacent to the unit being attacked". This implies that flanking is focused on the defending unit now, which would be a change. But as you just said that is not the case, it is still based on the attacking unit it would appear.

I believe you are saying it is still based on the attacking unit, but instead of "attacker gets all of its flanking bonuses and defender gets all of its flanking bonuses" its "only 1 side gets a bonus based on a delta".

2) I think we all understand that the flanking modifiers like shock are different now (static additions as opposed to multipliers). However, your statement above suggests that the core 15% flanking modifier did not work "normally". Is this correct...in the original mechanic, was the 15% flanking bonus not applied like a standard combat modifier?

Aka if I had a Drill I unit with 1 flank in the old system it "said" I had a +25% bonus applied to my units' strength (10 drill + 15 flank). Was this not how the math actually worked?

From what I remember, that is not actually correct on your last point with respect to the math. And the calculation got even more construed when adding more flankers.

I’d like to set up a few scenarios comparing old vs new when I finish work today.

 

[Gazebo]

Little questions.

1. Is the combat algorithm able to move several units, fire, then move them back within the turn? I mean, things like moving a pathfinder to provide flanking to a melee unit, let the melee unit attack, then retreat the pathfinder.
2. Is the combat algorithm able to reuse within the same turn a unit that were set to attack a threat that has already been dealt with? Like, a spearman that was set to attack an archer, but this archer has been killed by another unit, so the spearman is free to attack another unit in the same turn?
3. Is the combat algorith able to move great generals/admirals before the shooting phase begins, and move it again to another hot point after the combat in the former point is over?
4. Is there a combat stance for crippling units rather than going for the kill? In some cases (comparing my attack/defense power in the zone vs the attack/defense power of the enemy) I may prefer to cripple enemy units so they can't attack at full power and are forced to retreat, so I don't have to face fresh units. (I know many players would think that the AI is not working properly if it didn't go for the kill, but sometimes it is just better to not kill first).
5. How does the combat algoritm deal with incoming pillagers (fast moving units, generally alone, that attack from the rear hoping to pillage as much as posible)?

1. Somewhat - the biggest issue for the AI is imparting the knowledge that vision provides within a turn. Because you run into two issues: an AI that cheats by 'remembering' what it might see if it moves a unit somewhere or an AI that completely rebuilds its plans after every move. The latter is just too slow.

2. Once a unit is set as 'processed' it is removed from the potential pool of unit actions.

3. Yes, but generally the AI avoids this because it might result in a GG/GA being in a danger spot or being isolated.

4. Yes, the AI will try to 'attrit from range' when defending against a superior force or attacking a superior force. The idea here is to screen with melee enough to let ranged units pick away at targets. It is about as effective in AI hands as in human hands, except that it's hard for the AI to find the sweet spot for 'bailing out' and retreating or advancing.

5. Danger plots will get assigned, nearby units will move to intercept.

G, with respect, the reason that we are all "overthinking it" is because we have been given very little real understanding of the new mechanic. You just throw some change notes on us and left us to fend for ourselves

1) The change notes say "the bonus itself was reduced to 5%x the delta of friendly v. unfriendly units adjacent to the unit being attacked". This implies that flanking is focused on the defending unit now, which would be a change. But as you just said that is not the case, it is still based on the attacking unit it would appear.

I believe you are saying it is still based on the attacking unit, but instead of "attacker gets all of its flanking bonuses and defender gets all of its flanking bonuses" its "only 1 side gets a bonus based on a delta".

2) I think we all understand that the flanking modifiers like shock are different now (static additions as opposed to multipliers). However, your statement above suggests that the core 15% flanking modifier did not work "normally". Is this correct...in the original mechanic, was the 15% flanking bonus not applied like a standard combat modifier?

Aka if I had a Drill I unit with 1 flank in the old system it "said" I had a +25% bonus applied to my units' strength (10 drill + 15 flank). Was this not how the math actually worked?

That's because I literally threw the new version up an hour before I had to get on a train to leave. I just didn't have the time.

1. Attacker-based, but if the # of enemy units exceeds friendly units adjacent, there is no bonus. Formerly there would be (it didn't care nor check for enemy unit numbers).

2. Correct - it was 15% more flanking damage, but not 15% more damage. So if you were getting, say, a 20% flanking bonus, you'd be getting 15% more of that (i.e. +3 CS), not 15% more period.

G

 

[tu_79]

1. Somewhat - the biggest issue for the AI is imparting the knowledge that vision provides within a turn. Because you run into two issues: an AI that cheats by 'remembering' what it might see if it moves a unit somewhere or an AI that completely rebuilds its plans after every move. The latter is just too slow.

2. Once a unit is set as 'processed' it is removed from the potential pool of unit actions.

3. Yes, but generally the AI avoids this because it might result in a GG/GA being in a danger spot or being isolated.

4. Yes, the AI will try to 'attrit from range' when defending against a superior force or attacking a superior force. The idea here is to screen with melee enough to let ranged units pick away at targets. It is about as effective in AI hands as in human hands, except that it's hard for the AI to find the sweet spot for 'bailing out' and retreating or advancing.

5. Danger plots will get assigned, nearby units will move to intercept.

Wow. AI is miles better than I thought it to be!

So, the only room for improvement without cheating (or seeing more than it is available) is releasing half-used units to be used again in nearby danger plots. This is something I do with units that provided flanking for a kill, but did not expend all of its movement. They might provide flanking against other units.

 

[Stalker0]

Notes in the text.

That's because I literally threw the new version up an hour before I had to get on a train to leave. I just didn't have the time.

—hey we all need time away I get it, just understand that’s why we got some of the uproar in the thread. But we will get it all straightened out.

1. Attacker-based, but if the # of enemy units exceeds friendly units adjacent, there is no bonus. Formerly there would be (it didn't care nor check for enemy unit numbers).

—so the bonus ranges from 0 to X. The defender will never get a bonus like in the old system.

2. Correct - it was 15% more flanking damage, but not 15% more damage. So if you were getting, say, a 20% flanking bonus, you'd be getting 15% more of that (i.e. +3 CS), not 15% more period.

—let me try an example. Let’s use the example of 1,2,or 3 units flanking.

1 unit: +15% cs (from the 15% flanking bonus).

2 unit: +17.25% cs (15% from the first one. The second is 15% of that 15%, or 2.25%. So overall it’s 17.25% cs)

3 unit: +19.8% (17.25 from the first 2, 15% of thst is 2.58. So added it’s 19.8)

Is that right?

G

 

[Questdog]

Notes in the text.

Notes in the text style, makes it hard for anyone to quote YOU......


"1. Attacker-based, but if the # of enemy units exceeds friendly units adjacent, there is no bonus. Formerly there would be (it didn't care nor check for enemy unit numbers).

—so the bonus ranges from 0 to X. The defender will never get a bonus like in the old system."

This is wrong; the defender will get a bonus if he has more units flanking the attacker than the attacker has flanking him.

The difference in the new system, is that the offensive and defensive flanking bonuses can cancel each other out. In the old system, both would still get their flanking bonuses, making the calculation scary to the A.I. apparently, because he would see the defender's flanking bonus and shy away even if he had offensive flanking bonuses to compensate. The net result is the same either way, however.

Edit: Well, not exactly the same. Two 5% flanking bonuses canceling out is not the same as two units each getting 5% bonuses if one of the units has a much higher Combat Strength.

 

[Stalker0]

Notes in the text style, makes it hard for anyone to quote YOU......


"1. Attacker-based, but if the # of enemy units exceeds friendly units adjacent, there is no bonus. Formerly there would be (it didn't care nor check for enemy unit numbers).

—so the bonus ranges from 0 to X. The defender will never get a bonus like in the old system."

This is wrong; the defender will get a bonus if he has more units flanking the attacker than the attacker has flanking him.

The difference in the new system, is that the offensive and defensive flanking bonuses can cancel each other out. In the old system, both would still get their flanking bonuses, making the calculation scary to the A.I. apparently, because he would see the defender's flanking bonus and shy away even if he had offensive flanking bonuses to compensate. The net result is the same either way, however.

Edit: Well, not exactly the same. Two 5% flanking bonuses canceling out is not the same as two units each getting 5% bonuses if one of the units has a much higher Combat Strength.

Actually I canceled bonus is the same as giving the same bonus to both units… Because damage is determined by combat ratio, giving the same percentage bonus to both units maintains the ratio

 

[JohanBest]

What's the current status of the spearman line against ranged mounted units?

The code to check for mounted ranged units in combat has been removed this patch. So since the ranged mounted units are considered to be "UNITCOMBAT_ARCHER" type then any unit with bonus vs mounted won't get it against these units.

Not sure why this was removed. As of last patch it didn't work correctly (IMO) as spearmen etc shouldn't get a bonus in defense vs the ranged of horse units. I had a fix to this merged into the solution so that the bonus was given only to spearmen when attacking a ranged mounted unit. Dunno if my fix was intentionally left out of the final patch or not.

Oh well, looks like there are more pressing "issues" atm