Attack Delay...design philosophy

[Bsharp]

'Attack delay' was a HOI3 addition and I personally don't believe it has improved the game.

As I understand it, the logic is that (IRL) unit leadership requires some time to regroup/reorganize before a new attack can be launched. In game terms, attack delay limits the advance rate of armies to more historical pace.

As implemented, there are several problems:
- No allowance is made for the relative size/strength of the forces (one soldier can delay an army)
- A long, hard fought battle often ends with the attack delay timer expired and the attacker badly damaged but free to immediately re-attack; while a short walkover battle ends with the attacker at full strength but paralyzed for days waiting for the timer to count down.
- A division attacked while moving gets delay, but a division attacked at rest does not
- A division with an org level of 70 is delayed just as much as a division with an org level of 5.

I believe the designers should scrap 'attack delay' and address the 'historical rate of advance' issue in other ways.

If I remember correctly, HOI1 and 2 used the philosophy that movement, even when unopposed, causes a unit to become disorganized. Thus, a panzer division that blitzed across 200 km of open Ukrainian countryside would arrive at its destination with less fighting ability than it started. That seems perfectly logical to me, but it's exactly the opposite of the way HOI3 works. In HOI3, you recover organization as you move. Not only is that unrealistic, but I believe it contributes to the excessive advance rates that created the need for an attack delay.

Just as air units have returned to the HOI1 & 2 convention of losing org when changing bases, so should land units return to losing org when they move. And, then we wouldn't need the much-hated attack delay. Thoughts?

 

[Karmapowered]

Either that, or fine tune the current attack delay to take more parameters into account.

You already listed quite a few, among which the relative strength and organization of the opponents, but we probably also could mention the geographical configuration (terrain), the current weather, whether enemy air forces are active or not, etc.

Basically, all the factors that, after a battle, would slow down an army in returning to the horses/trucks/rail, to advance or retreat.

Basically however, I believe you did a fine job at summing the issue up

EDIT : Also, start the attack delay once the battle is over, yes. Whatever the real motivation was that could explain initiating the delay as soon as the battle starts, it seems fairly unintuitive to me.

'Attack delay' was a HOI3 addition and I personally don't believe it has improved the game.

As I understand it, the logic is that (IRL) unit leadership requires some time to regroup/reorganize before a new attack can be launched. In game terms, attack delay limits the advance rate of armies to more historical pace.

As implemented, there are several problems:
- No allowance is made for the relative size/strength of the forces (one soldier can delay an army)
- A long, hard fought battle often ends with the attack delay timer expired and the attacker badly damaged but free to immediately re-attack; while a short walkover battle ends with the attacker at full strength but paralyzed for days waiting for the timer to count down.
- A division attacked while moving gets delay, but a division attacked at rest does not
- A division with an org level of 70 is delayed just as much as a division with an org level of 5.

I believe the designers should scrap 'attack delay' and address the 'historical rate of advance' issue in other ways.

If I remember correctly, HOI1 and 2 used the philosophy that movement, even when unopposed, causes a unit to become disorganized. Thus, a panzer division that blitzed across 200 km of open Ukrainian countryside would arrive at its destination with less fighting ability than it started. That seems perfectly logical to me, but it's exactly the opposite of the way HOI3 works. In HOI3, you recover organization as you move. Not only is that unrealistic, but I believe it contributes to the excessive advance rates that created the need for an attack delay.

Just as air units have returned to the HOI1 & 2 convention of losing org when changing bases, so should land units return to losing org when they move. And, then we wouldn't need the much-hated attack delay. Thoughts?

 

[Alex_brunius]

I can only fully agree.

I believe it is being looked over though. I think a recent change was to prevent units moving into hostile territory gaining org like they did in earlier versions of HoI3. Not sure if they lose any org though, I do know paras lose a bit off org when they drop now.

 

[zeekater]

I always figured that the attack delay simulated armies like the French with their WW1 style communications which meant that there was so much time between the generals knowing what is going on, giving orders and the troops receiving those orders.

Having units lose org while attacking hostile territory (even unopposed) would be nice

 

[Alex_brunius]

While we are talking about org regain, I wonder why they removed the diminishing returns off org regain too (In HoI2 your units gained org very fast if close to 0, and very slow when close to max).

That mechanic meant that defeated units had a realistic chance of stopping advancing armies even if they have to lose several battles and alot off strength in the process.

 

[zeekater]

While we are talking about org regain, I wonder why they removed the diminishing returns off org regain too (In HoI2 your units gained org very fast if close to 0, and very slow when close to max).

That mechanic meant that defeated units had a realistic chance of stopping advancing armies even if they have to lose several battles and alot off strength in the process.

Good point, now you rarily see a division with low strength (unless you bomb one to smithereens).

 

[unmerged(139259)]

Good points here, but I think only three things would be quite enough to leave defender some chance to regroup and maybe even recostruct the frontline - which would finally mean better balancing and a need for real phasing of offensives, instead of constant steamrolling we see now.

1. Steady minor org loss for units advancing into enemy territory, not applied when moving on their own territory (i.e. controlled). This would prevent unlimited raids deep into enemy territory and would make retreats much more reasonable, as the attacker would have to make his way through our rear echelons before reaching our main defensive line, which would somehow weaken him (that's why org loss).

2. Decrease unit supply stockpiles from 30 to only 10 or 20 days (which would actually force attacking units to stop from time to time to let the supplies catch up on them) and make infra levels much more important by adding severe penalties to movement speed through provinces with bad infra.

3. Start counting attack delay from the END of the battle, not from the start of the battle. Moreover, the length of attack delay should be deduced from the length of the battle - the longer battle, the longer delay (up to the max amount set by current tech). From the game perspective, this would also help defenders to get enough time to recover from tough and lengthy battles, while the short ones would not stop attackers for days.

From the real life perspective, I know that it should reflect delays in command chains, but when you think of it, this does not work as it should. After all, HQ can give additional orders only according to how the battle has developed in the end, not that it would be able to predict it. Currently, attack delay simulates a situation like this - Adjutant: "Our unit XY reports that it has just commenced an attack on province A, sir." Commander: "Well done. Now we should conquer province B." Adjutant: "Sir, but the attack has just begun. We haven't achieved the necessary breakthrough yet." Commander: "Never mind, when my orders reach the unit XY after 72 hours, the breakthrough will surely be achieved already." Adjutant: "But how do you know, sir? Shouldn't we react rather to how the battle evolves?" Commander: "I see it in my crystal ball, so I'm issuing any command immediately after another. After all, that's how the engine calculates the attack delay." Adjutant: "What's the attack delay, sir?" Commander: "Noone knows, but the rumours say it's a magic tool that allows a single soldier to stop an entire army for days just sitting on the spot doing nothing. Like the Italian Army in Western Desert."

 

[unmerged(219522)]

Perhaps making attack delay proportional in some way to time spent in battle could give a more realistic result. Even this system would not be perfect though, for example German tank divisions after winning a battle which would force a breakthrough would not spend time planning on where/how to advance next; they would simply drive on and exploit that victory. Maybe somehow there should also be the number of troops in the "next" province taken into account...Overall a system of "attack delay" will never be perfect in a game of this system as there were many different factors which would effect this, rather than generic across the board.

 

[podcat]

While we are talking about org regain, I wonder why they removed the diminishing returns off org regain too (In HoI2 your units gained org very fast if close to 0, and very slow when close to max).

that was added back in one of the SF patches if I dont completely have things mixed up.

 

[Sirveri]

3. Start counting attack delay from the END of the battle, not from the start of the battle. Moreover, the length of attack delay should be deduced from the length of the battle - the longer battle, the longer delay (up to the max amount set by current tech). From the game perspective, this would also help defenders to get enough time to recover from tough and lengthy battles, while the short ones would not stop attackers for days.

Same thing, but worded differently.
Start counting up to the cap on the timer when battle starts.
When battle stops start counting back down.

So for example you get in a fight you start the counter at 0 hours, then the fight rages for 3 hours because you're just breaking a HQ unit, then it starts counting down once the battle finishes. Then you fight a real battle and it takes 170 hours, however you have one level of OLO so your cap is 144, so when battle finishes it starts counting down from 144. It's the same thing, but it's easier to do from a programming perspective since it's just a simple counting loop.

int cTimer(int cap, bool combat)
{
  int i = 0;
  while(combat)
  {
    i++;
    if(i==cap)break;
  }
  return i;
}

Think my code is a bit rusty but that might be a functional function.

 

[Raptor83]

that was added back in one of the SF patches if I dont completely have things mixed up.

Its not currently implemented or effect is so small that I dont see it.

 

[Bullfrog]

Org loss itself simulates the same thing as "attack delay" so I see little reasoning behind it. If org loss was modified to be higher, attack delay could be done away with entirely.

 

[jackda]

Its not currently implemented or effect is so small that I dont see it.

From defines.lua

LOW_ORG_REGAIN_BONUS = 0.30,

You can change this value for a higher one if you feel the need.

 

[juv95hrn]

Its not currently implemented or effect is so small that I dont see it.

It's notified as implemented and the effect was described as quite drastic.

Personally I liked that units lost some org in HOI2 when entering hostile provinces.

 

[unmerged(304296)]

Good points here, but I think only three things would be quite enough to leave defender some chance to regroup and maybe even recostruct the frontline - which would finally mean better balancing and a need for real phasing of offensives, instead of constant steamrolling we see now.

1. Steady minor org loss for units advancing into enemy territory, not applied when moving on their own territory (i.e. controlled). This would prevent unlimited raids deep into enemy territory and would make retreats much more reasonable, as the attacker would have to make his way through our rear echelons before reaching our main defensive line, which would somehow weaken him (that's why org loss).

2. Decrease unit supply stockpiles from 30 to only 10 or 20 days (which would actually force attacking units to stop from time to time to let the supplies catch up on them) and make infra levels much more important by adding severe penalties to movement speed through provinces with bad infra.

3. Start counting attack delay from the END of the battle, not from the start of the battle. Moreover, the length of attack delay should be deduced from the length of the battle - the longer battle, the longer delay (up to the max amount set by current tech). From the game perspective, this would also help defenders to get enough time to recover from tough and lengthy battles, while the short ones would not stop attackers for days.

From the real life perspective, I know that it should reflect delays in command chains, but when you think of it, this does not work as it should. After all, HQ can give additional orders only according to how the battle has developed in the end, not that it would be able to predict it. Currently, attack delay simulates a situation like this - Adjutant: "Our unit XY reports that it has just commenced an attack on province A, sir." Commander: "Well done. Now we should conquer province B." Adjutant: "Sir, but the attack has just begun. We haven't achieved the necessary breakthrough yet." Commander: "Never mind, when my orders reach the unit XY after 72 hours, the breakthrough will surely be achieved already." Adjutant: "But how do you know, sir? Shouldn't we react rather to how the battle evolves?" Commander: "I see it in my crystal ball, so I'm issuing any command immediately after another. After all, that's how the engine calculates the attack delay." Adjutant: "What's the attack delay, sir?" Commander: "Noone knows, but the rumours say it's a magic tool that allows a single soldier to stop an entire army for days just sitting on the spot doing nothing. Like the Italian Army in Western Desert."

1) Do not understand the reasoning behind no ORG loss for movement within own country. Movement should cause ORG loss, combat should cause higher ORG loss.

2) 30 day supply is required because if the supply mechanics. Lowering this will result in units running out of supply even more erratically than before. This has been discussed in other threads.

3) This would require redoing Attack Delay completely. The current system cannot be modified in this way. Again, discussed in other thread.

 

[DerKomtur]

> and make infra levels much more important by adding severe penalties to movement speed through provinces with bad infra.
I'm pretty sure there already are severe penalties to movement for bad infra.

About org loss on the move: Instead of an unconditional loss on the move, let org of a unit tend toward 50% (or whatever modable value) of it's maximum org. When a unit is very low on org, it gains org while moving, while a unit with full org will lose org.

I strongly support a redesign of the attack delay. Especially the fact that after a one hour skirmish you need to rest a week; while after a long battle the attacker can immediately continue attacking which is exacerbated by the fact the attacker has already finished his march to the contested province, but the defender is just starting to retreat.

 

[hito1]

(...)
About org loss on the move: Instead of an unconditional loss on the move, let org of a unit tend toward 50% (or whatever modable value) of it's maximum org. When a unit is very low on org, it gains org while moving, while a unit with full org will lose org.
(...)

Very elegant solution!

 

[unmerged(304296)]

Thinking about it Attack Delay is the wrong way to handle this. All that is needed is to make attacking dependent upon ORG level in the same way PARA drops are (which are themselves a special form of attack). The tech could be changed so each level decreases the %ORG needed to attack. Officer % over 100% also would decrease it. This is actually an excellent model of how it works in R/L. When an attacking unit is too disorganized by the combat it needs to pull itself together before it can go on.

 

[truth is life]

Thinking about it Attack Delay is the wrong way to handle this. All that is needed is to make attacking dependent upon ORG level in the same way PARA drops are (which are themselves a special form of attack). The tech could be changed so each level decreases the %ORG needed to attack. Officer % over 100% also would decrease it. This is actually an excellent model of how it works in R/L. When an attacking unit is too disorganized by the combat it needs to pull itself together before it can go on.

You know, I suggested this exact system a loooong time ago, and I still think it's as valid an idea today as it was then.

 

[unmerged(304296)]

You know, I suggested this exact system a loooong time ago, and I still think it's as valid an idea today as it was then.

The best idea will always be valid.