Change back the dog fight mechanics

It took a second cup of coffee this morning, but I think I understand.

With multiple groups in the same area you want to be the defender, not the attacker. And by "attack" I mean patrol tick here in this post.

To make it simpler lets deal with two stacks of planes on each side, all in one patrol area. I am going to use level 3 planes, so interceptors do 5.5 damage both ways, Tac bombers do 1.5 offense and 3 defense. For balanced stacks of 5 interceptors and 5 tac bombers they will do the following: On patrol tick 8.75 , and defending 10.5. Interceptors have 20 hit points each, and tac bombers have 25 hit points each, for a total of 225 hp per stack. In reality the x factor will alter this somewhat.

Enemy group #1 reaches their patrol tick and dings both of my groups. 8.75 hp each spread among 10 planes leaves my stacks at 96%. Now Both of my groups return fire on enemy group #1, so it takes 21 damage at once, so the enemy group loses the equivalent of one interceptor, or the group is left at 90%.

If we let enemy group #2 take their patrol tick we should end up with all groups around 90%, And then allowing my groups to do their patrol tick the numbers should balance out.

Now if we go with the Roko/meich scenario of 5 groups of planes on each side:

Enemy group #1 reaches their patrol tick, and fires on each of my 5 groups of planes, so I now have 5 groups of planes at 96%. Each of my 5 groups of planes returns fire for a damage of 52.5 hp to enemy group #1, or a loss of 2 planes. So the enemy is taking big losses in order to scratch the paint on my aircraft.

By the time each of the 5 enemy stacks has taken their patrol tick I have absorbed 43.75 damage on each of my 5 stacks, and dealt 52.5 damage to each enemy stack. So I have half a plane more left. It does seem alarming to lose a couple planes on the first patrol tick, but it should balance out in the end, unless I am missing something. Equal numbers at the beginning mean equal losses by the end. I fail to see an exploit here with balanced numbers. Now if the enemy has one large group on patrol and I move in with 5 small groups that might be something, but I will have to take time to run the numbers.

The only exploit on my part would be if the enemy left planes on patrol over their formation of ground troops including antiair. I could move in with several groups of planes and let them get hit by the patrol tick and eliminate the enemy planes, but not wait for a patrol tick of my own so that I would not get hit by both the planes and the antiair.

miech wrote:

Ive seen disproportional losses, even with the defensive bonus. Think getting 11 planes killed (out of 20-25, cant remember clearly) kind of damage during one of my offensive ticks, nearly no damage done back (think 0-5% per fighter)

freezy wrote:

by making the offensive tick trigger right at the start of the patrol timer and not after 15min,

Maybe it is easier to implement that units can't shoot as defenders against an enemy tick, when they have already done so in the last 15m?

freezy wrote:

... with multiple stacks vs less stacks to trigger the exploit, ...

K.Rokossovski wrote:

freezy wrote:

by making the offensive tick trigger right at the start of the patrol timer and not after 15min,

That sounds like a horrible solution in non-air-opposed situations... when it ticks at the start, you can just shoot-reposition-shoot-reposition many times in a minute? Or if you can't, it doesn't solve anything?

Maybe it is easier to implement that units can't shoot as defenders against an enemy tick, when they have already done so in the last 15m?

So while we are at it we are still open for better suggestions. if they are not too complicated, you can post them here, I will talk them through with the devs to see if they are feasible.

Not sure if I understand your idea correctly, because disabling defense attacks for 15min after your last defensive attack would also not solve the problem, as your enemy only attacks once in 15min anyway. It would slow down fights though if multiple stacks are fighting on both sides, maybe evening out the combat a bit more, but I am not sure. Maybe someone can do an example calculation with this? Also we have to think about some edge cases where this might backfire somehow.

K.Rokossovski wrote:

freezy wrote:

... with multiple stacks vs less stacks to trigger the exploit, ...

it has nothing to do with the difference in amount of stacks, just that there is more than one stack defending. If there are 5 attacking stacks against 2 defending stacks, on their own tick, those five stacks all shoot once for 5X damage, while the defenders all shoot twice on each tick for 10X damage.

If we remove defensive damage again, then the balancing problem applies for the split offensive stacks.

freezy wrote:

To fix this we would need to rewrite the combat system to apply the stack penalties after the damage of each stack is combined, which is sadly easier said than done as the combat system is old and spread in many lines of code at many different places. Maybe that can be done in the future, but right now we don't have the capacity for that.

I've already posted the fix for this problem last year. It's fairly easy to implement, but the staff at the time got upset with me for using too much flamboyance in my presentation and my zeal for the solution was such that I got in trouble. However, the solution itself is still valid and I suspect that none of the dev team ever saw it because I was shut down in the process. But, since it's relevant again -- and more so, now -- I present to you MY SOLUTION to this problem...as found in SBDE - New Feature

Make ALL units in overlapping patrol zones over the target adjust their SBDE such that they fight as if they are in one single stack of units.

This way, though one patrol attacks five other patrols and they all shoot back, they must act as if they are one big group for SBDE consideration. Obviously, the same must be true of your own units on offense. When one of your groups attacks on patrol, every enemy unit/stack within your patrol radius that is ALSO within the patrol radius of other units within your circle will have your attack reduced by your own combined SBDE such that neither side gets an exploit.

Now, if this is implemented, it also solves the SBDE exploit for which that other post was originally trying to solve. That other exploit is the exploit over ground forces where overlapping smaller patrols avoid the SBDE penalties without forcibly merging whereas their targeted ground forces are always required to merge within range of each other which has the consequential effect of lowering their SBDE.

They're both simple exploits, but fixing either of them automatically fixes the other and my solution works beautifully to solve both. Just be sure to check out the original post which more clearly explains the fix. And, in addition to that post, if you scroll to the top of that thread, you'll find an older idea of mine for increasing SBDE capabilities for units based on combining various unit types together (as opposed to all of one unit type).

That other -- older -- idea, though a bit more involved, might be fairer all around such that everyone's happy. Besides, I know more than one player that would likely kill me if my more recent idea to fix the SBDE exploit for Tac Bombers were to get implemented...but that's another story and another complaint that was covered in a couple other threads a long time ago.

freezy wrote:

So while we are at it we are still open for better suggestions. if they are not too complicated, you can post them here, I will talk them through with the devs to see if they are feasible.

Well, if Bytro can implement one of my ideas as listed above, I think the exploit/bug would not only be fixed but the game would improve substantially for most players. I hope you can review my ideas and I believe that both are feasible and either one would help the situation. However, the first idea I listed above would go a lot further in fixing the problem for this latest exploit that @miech initially described.

freezy wrote:

the combat system is old and spread in many lines of code at many different places.

Dang it. This is just so typical of older code when it gets edited by numerous employees, especially as different ones leave and new ones come into a project over a long period of time. This is a classic source of "spaghetti code" and it makes it very difficult for new coders to take ownership of existing code for which they become responsible. Besides, even those very familiar with the project suffer from the dis-information leak of spaghetti code over time.

I've always espoused a strict adherence to coding standards and documenting changes in the code. Among these includes: a proper separation of the implementation from definition, reduction of methods to minimal functionality, and hyper-specific interface protocols while keeping changes to an absolute minimum. It is so much easier to add to a small function with a wrapper than to change the function itself.

The encapsulation principles of this OOD methodology are so much cleaner and they make error checking and whitebox testing so much more efficient. Plus, well-defined project reports, change logs, updated design specs, thorough inline documentation, and periodic review of the code changes against lower-tier [older] revisions in an ordered repository can do much to keep everyone who is a coding-stakeholder familiar with the code irrespective of who comes into a project, even years after it is first designed. And this old-school [read "well-documented'] way of making code changes can work well within an Agile coding environment so long as those standards are enforced despite the smaller and more fluid progression segments of an Agile implementation.

Diab, I think your solution creates more new problems than it solves... for example, what happens when patrol circles are not the same between air groups? And there are other problems with it as well; like SBDE won't solve all the problems when air groups are taking losses during the battle.

@freezy, IMHO, the following adjustments would solve most of the current air war troubles and also give less-active-than-24/7 players a bit better chance in air battles:

- Patrolling is a defensive thing only; they don't have a "tick" but respond to enemy air-to-ground (or air-to-sea) action in their patrol circle only;

- If you want to attack ground troops or ships from the air, you HAVE TO use direct attack (AND run the refuelling risk afterwards);

- When you want to attack an enemy (patrolling or moving) AIR group, you should also use direct attack. This type of combat already exists, but should be rebalanced, because currently it is disastrous for the attacker in most cases;

- Enemy patrolling planes do NOT interfere with direct air-to-air attacks, even in their patrol circle (to avoid the same problem persisting because direct A2A attack would trigger defensive shots from all patrolling air groups)

Oh and diab, I think a software development discussion is out of scope for this forum... but if I was adressed with these "open doors" for my own software development, coming from someone that has no idea about my work procedures or methodology, I would be quite insulted and likely ignore it completely. I'm sure Bytro uses a fine methodology, or if they don't, are working very hard to adress the problems without any need to be told about general software development practices.

what if patrolling planes took an unanswered attack when they broke/changed patrol. the exploit is that folks break patrol after 14 minutes. just make it so a plane takes an extra attack from patrolling enemy planes as they retreat from their own patrol. then the exploit is defeated.

Diabolical wrote:

Make ALL units in overlapping patrol zones over the target adjust their SBDE such that they fight as if they are in one single stack of units.

And yes, you say it, the combat mechanic is pretty much spaghetti code. Nowadays we don't code new features in that way anymore and adhere to coding standards you also listed. But you have to keep in mind that the combat mechanic in CoW is based on S1914, and that is ~10 years old. Back then the code just wasn't written with these standards in mind

K.Rokossovski wrote:

- Patrolling is a defensive thing only; they don't have a "tick" but respond to enemy air-to-ground (or air-to-sea) action in their patrol circle only;

- If you want to attack ground troops or ships from the air, you HAVE TO use direct attack (AND run the refuelling risk afterwards);

- When you want to attack an enemy (patrolling or moving) AIR group, you should also use direct attack. This type of combat already exists, but should be rebalanced, because currently it is disastrous for the attacker in most cases;

- Enemy patrolling planes do NOT interfere with direct air-to-air attacks, even in their patrol circle (to avoid the same problem persisting because direct A2A attack would trigger defensive shots from all patrolling air groups)

Stormbringer50 wrote:

what if patrolling planes took an unanswered attack when they broke/changed patrol. the exploit is that folks break patrol after 14 minutes. just make it so a plane takes an extra attack from patrolling enemy planes as they retreat from their own patrol. then the exploit is defeated.

fix the unlimited Patrol time!

in CoN, planes on patrol that engage/are engaged by ANY enemy unit immediately thereafter RTB.

pretty simple.

freezy wrote:

This might by possible to do according to the Devs. But it of course changes alot how air combat works. Would everyone be satisfied with such a drastic change? Patrol would then be a clear defensive mechanic only, and attack the only offensive mechanic for planes. Post your opinions here

Treating all stacks as one big pseduo stack would take away the advantage. I think there is currently too much of an advantage for the person that is online and interactively controlling their planes, but I think there should be some reward for those willing to put in the effort. I would suggest having the patrol randomly pick two air stacks and attack them instead of all the stacks. It would still give the online player some advantage, but more like 2x instead of UNLIMITED-x. Since the code must already have to iterate over stacks in range this should not be hard to implement. I also do not think ground stacks should ever be combined. Planes already have enough of an advantage over ground.

freezy wrote:

And yes, you say it, the combat mechanic is pretty much spaghetti code. Nowadays we don't code new features in that way anymore and adhere to coding standards you also listed. But you have to keep in mind that the combat mechanic in CoW is based on S1914, and that is ~10 years old. Back then the code just wasn't written with these standards in mind

It's just one of those things where you have to decide when it's time to stop the bleeding by having an arterial surgery...yes, you'll bleed a lot during the surgery, but the long term benefit is improved health...or improved pocketbooks.

Stormbringer50 wrote:

freezy wrote:

This might by possible to do according to the Devs. But it of course changes alot how air combat works. Would everyone be satisfied with such a drastic change? Patrol would then be a clear defensive mechanic only, and attack the only offensive mechanic for planes. Post your opinions here

imo, absolutely not. The answer to an exploit would be to drastically nerf the unit? In all honesty, as a great fan of air power, I would fear that I would lose all interest in the game.

jk

DxC wrote:

Treating all stacks as one big pseduo stack would take away the advantage.

Well, that's kinda my point. I want to remove the advantage because it's basically an exploit. If you read some of my old posts on this, you'll see where I'm coming from and how I came to realize just how bad of an exploit it is. Note, that these are all closed threads, so you can't add to them, but if you have comments about any of them, I think this thread is the appropriate place to put those at this time.

The One True Change that Must Happen

Tactical Bomber Problem - Revolutionary Solution

SBDE - New Feature

I understand the so called exploit and the consequence of treating everything as one stack. I'm just disagreeing with this solution, even if Bytro were in a position to change it.

someone just had to get diabolical started. diab, they already raised the aa capabilities of everything with aa in the game. what more do u want?

and as far as leaving the game, that is not what i said at all.

Other idea to stop this situation is to disable the unlimited new positioning for planes in the "patrolling rectangle" (or pie) for all air units.

If all air units must go back to air base after 3 new flight points we fix this problem easily.

But I prefer the idea to limit the total patrolling time to 1 hour (4 ticks) after this time the planes must go back to air base and need to refuel. If the devs do it in this way we lower a bit the power of air unit in this game.

sure, as long as we make armored units return to a home province to refuel (or wait for supply trucks), make naval units return to return to port to refuel, make infantry units have to rest after they march for 8 hrs. Arty should have to rest and not be able to fire after 3 rounds, as they wait for ammunition to be brought to them. during these rest periods, these units should be in a weakened condition, like refueling planes.

anyone get my point?

The answer here is still not to nerf the unit.

if u have 3 stacks of arty, or naval stacks bombarding from 3 different directions, should u drop their efficiency because they are all attacking the same target?

this was a post about an exploit/bug that players have turned into an opportunity to try to ruin aircraft.

I might have a simple solution here:

The exploit revolves around the fact that planes moving from one patrol location to another are in a straight line flight mode instead of a patrol.

Would it be possible that once a patrol is established that the patrol circle and timer will be active until the planes either attack or land at an airfield? You could still re position planes in their green slice of pie, but as they are moving they still have the patrol circle and the timer still ticks.

Under these rules:

1) The only way to end a patrol is to either attack, or refuel.

2) Once an attack order is given, the patrol option is not available until after refueling, even if the attack is cancelled

Now as for speed of the patrol moving I wonder if it should be slower, perhaps 70% of the straight line speed, although that would interfere with returning to base while providing air support to planes that did an attack run.

Opinions?

it seems yo me, imho, that all these suggestions require major reprogramming of flight mechanics. that being said, the simplest solution still seems to me to be: If a patrol is broken to reposition, let enemy aircradt in the patrol circle and any unit with AA have a free unanswered attack on them. this kills the exploit, doesnt nerf the unit, and actually makes sense as the retreating aircraft are fired upon as they regroup their flights to reposition their patrol circle.

DxC wrote:

Vorlon. You no longer have to be moving so all of what you just wrote is irrelevant.

Is this supposed to be funny or only spam?