Hi Elmo,
The AI in AB is certainly not 'fully autonomous' as it requires constant input from the player in the form of orders.
The point of contention is whether you are giving those orders based on a god like perspective of the battlefield or one that incorporates information delays.
In both cases you're still issuing orders but you doing so with either perfect information or imperfect.
God like, perfect player information, appears to be the way the dev's are heading. Maybe the majority of people who buy a military sim like AB prefer it that way?
It could also be that the hassle of implementation outweighs the benefits.
Hard to say from the outside looking in.
Cheers,
Lancer
“Information delay” implementation
RE: “Information delay” implementation
We do have plans for improving the AI of individual units. They could have 'defensive' and 'aggressive' stances that can be selected in the SOP, and then they can use their initiative on their own, without the player intervening that much, hopefully increasing their chances to survive.
However, I think much more would be needed. Traffic jams in choke points would be frustrating to solve when the player doesn't even know where his units are. Pathfinding in general becomes more complex if we want to keep it realistic and every unit has only his own knowledge of minefields and other obstacles. So far our plan has been to keep that information global for the side.
I believe many (or most?) wargamers want to have a full control over their units, even if a fraction is talking about having limited friendly side information. Would it be any fun to play? Certainly it would confuse many players and make the game harder to approach. In the perfect world it would be included as an optional way to play, of course, but I'm not convinced it's worth the extra effort.
However, I think much more would be needed. Traffic jams in choke points would be frustrating to solve when the player doesn't even know where his units are. Pathfinding in general becomes more complex if we want to keep it realistic and every unit has only his own knowledge of minefields and other obstacles. So far our plan has been to keep that information global for the side.
I believe many (or most?) wargamers want to have a full control over their units, even if a fraction is talking about having limited friendly side information. Would it be any fun to play? Certainly it would confuse many players and make the game harder to approach. In the perfect world it would be included as an optional way to play, of course, but I'm not convinced it's worth the extra effort.
Know thyself!
-
Elmo_Zumwalt
- Posts: 17
- Joined: Fri Aug 24, 2018 10:52 pm
RE: “Information delay” implementation
That makes sense, and I agree that the return on investment is likely to be negligible, if not negative, since only a small portion of players would be interested in this feature. Ultimately this is a game, and capturing every aspect of reality is not possible (nor should that be the goal).
For instance, an actual commander would have a dedicated combat staff, HQ infrastructure, etc. to synthesize and filter information, which the player in AB does not. This factor alone changes the game considerably, as the player must constantly scan the battlefield themselves to see the critical points at any time.
I (and no doubt many others) would agree that improved, more autonomous AI would be a much more significant QOL improvement than less tangible features such as this.
For instance, an actual commander would have a dedicated combat staff, HQ infrastructure, etc. to synthesize and filter information, which the player in AB does not. This factor alone changes the game considerably, as the player must constantly scan the battlefield themselves to see the critical points at any time.
I (and no doubt many others) would agree that improved, more autonomous AI would be a much more significant QOL improvement than less tangible features such as this.
RE: “Information delay” implementation
Hi Veitikka,
We are talking at cross purposes here as I haven't explained myself very well.
I agree with you that the Player should have a god like view of his own forces for the reasons you've mentioned and full control.
What I was talking about, not very clearly, is having the disposition of the enemy forces shown on the map in a time delayed manner. For example a subordinate unit sees an enemy but the player receives that information after an appropriate time delay and what's shown on the map reflects the fact that the enemy position may not be current. The enemy unit icon is there but it's dimmed, etc.
The player is then making decisions and issuing orders based on the information about the enemy feeding up through the chain of command but at all times he knows exactly where his own forces are and has control over them in a normal manner.
Anyway, I'm sure whatever you do will work out fine, just thought I'd clear up the misunderstanding.
Cheers,
Lancer
We are talking at cross purposes here as I haven't explained myself very well.
I agree with you that the Player should have a god like view of his own forces for the reasons you've mentioned and full control.
What I was talking about, not very clearly, is having the disposition of the enemy forces shown on the map in a time delayed manner. For example a subordinate unit sees an enemy but the player receives that information after an appropriate time delay and what's shown on the map reflects the fact that the enemy position may not be current. The enemy unit icon is there but it's dimmed, etc.
The player is then making decisions and issuing orders based on the information about the enemy feeding up through the chain of command but at all times he knows exactly where his own forces are and has control over them in a normal manner.
Anyway, I'm sure whatever you do will work out fine, just thought I'd clear up the misunderstanding.
Cheers,
Lancer
-
nikolas93TS
- Posts: 703
- Joined: Fri Feb 24, 2017 4:32 pm
- Contact:
RE: “Information delay” implementation
Oh, that is actually a very cool idea, but I am not sure how implementable it would be. For example player would still see his units engaging "unseen" enemy, and the game might require a much more robust semi-autonomous friendly AI (to the point of maybe even overriding player commands in emergency).
Armored Brigade Database Specialist
RE: “Information delay” implementation
Hi Nikolas,
The player seeing the enemy engaging 'unseen' enemy units is exactly how the helicopters work in the game at present.
The friendly AI wouldn't need to change (although any improvement would be welcome). The game and friendly A.I would work as it currently does with the only difference is that the player is issuing orders based on the map situation as they know it which may well be different to how one of their subordinate units sees things.
Here's an example that highlights the good and bad of the idea and the current implementation of how things work.
Say your platoon of APCs spots enemy armour ahead. They send that information up the chain of command to you, the player, at HQ. Because of the delay in pushing the information upwards and perhaps some interference from EW, the information arrives 15 seconds later (real time). In the meantime, you issue an order for the APC's to advance, unaware of the armour threat.
The map doesn't yet show the enemy armour because you haven't received the information. Your APC's know that the enemy armour is there in front of them but until their sit-rep update arrives at your HQ, nothing is shown on the map.
So now you've issued an order that is quite possibly suicidal for your APCs. What happens? Well there is an order delay before implementation and you may issue another order once your recieve the update and the map shows the armour. Either way there is going to be a measure of chaos as you APC's advance a bit then reverse.
The other possibility is that your APC's may have preset operational parameters (like those that can be set by the player and already exist) that give them a measure of self preservation, regardless of your orders. This opens up another discussion about what level of autonomy should your subordinate units possess? Should they do exactly as you tell them or be able to 'query' an order they might consider dangerous (and how would you code the limits of what's dangerous and not?)
Contrast the above with the game as it currently is. The player has godlike awareness of his own forces (no different) and your APC platoon is giving you instant updates of everything they see.
So the player looks at the map, see nothing in the vicinity of the APC platoon and issues an order for them to advance. Shortly after doing so the enemy armour appears in front of them. The player sees this instantly (no delay in the information moving up through chain of command, no radio drop outs, no EW).
What's changed? Your APC platoon will still advance into the face of the enemy armour because you, the player, has issued an order to do so. You can issue another order to countermand the advance and have them reverse but you'll end up with the same chaos as above.
The APC platoon could have the ability to act in self preservation in accordance with some player determined operational doctrine but that is identical to the above and runs into the same issues of how much autonomy you would want your subordinate units to have.
The big difference between how the game currently is (all enemy activity shown instantly) and one where the map incorporates information delays on enemy activity is that the player will be able to react to having issued a bad set of orders faster (no information delay) with the current setup than with one which incorporates a delay. The friendly AI would act identically in both situations and would have the same issues.
So the question, perhaps, is what benefit would incorporating information delays have?
It'd make things more realistic. Whether it'd make it more fun is another question but if you were aiming the game such that a military audience was part of the deal it would be, I'd imagine, a big selling point.
It'd give you some scope to play around with things like EW and terrain in terms of their affect on information delay. I can't vouch for EW but I have a lot of experience in using radio equipment in various terrains and even high spec military equipment runs into 'drop outs' and 'low signal' once blocking terrain comes into play (eg. a hilly map). Sending units into difficult terrain and you're going to have issues with communications.
One aspect that the game doesn't model (and the competition does) is the amount of radio traffic inbound and outbound from an HQ. The enemy is actively searching for the high signal locations so they can drop artillery on them and hopefully take out HQ's. What you could do here is give the player an option for the level of traffic they are willing to live with. At low levels the information works it's way up the chain as described above but at high level perhaps it would represent the player at HQ issues lots of 'queries' to his subordinates, eg. 'what's going on? / are those tanks still there?, etc.' and this could be represented by a global reduction in the speed in which information travels up the chain of command.
You could tweak this further by having a global modifier to information delays that reduce them when the player's HQ unit is stationary and increase them when it's mobile. So a player then has to weigh up the benefits of improved communications and reduced information delays vs. the risk of being identified by the enemy signals surveillance and being targeted by an artillery strike. There's some potential for interesting and fun decisions here with the player having to hop his HQ around interspersed with short periods of stationary, high comms, activity when he really needs to know what's going on. What you would be doing, in effect, is giving the player a measure of control over their command cycle with a risk/reward game mechanic.
Lots of stuff you could do here with information delays to, perhaps, nudge the game towards a more interesting and realistic direction.
The big downside would be that it would highlight any friendly A.I issues even more so than currently as the player would not be able to react as quickly as they currently can with instant, on-map, enemy activity on the map. As highlighted above, there would be no difference in the reliance that the player would have on the friendly A.I but information delays would serve to bring any deficiences into a sharper focus.
The friendly A.I is already an issue (I can live with it but a lot of other people find it a concern) that, I'd suggest, you're going to have to address in some manner with the next version. It's an issue not because it's bad but because of how you've implemented the issuing of orders (with a delay) that requires friendly units to act autonomously and in a perceived sensible and logical manner during the intervals when you can't tell them what to do.
Whether putting another delay into the game mechanics around the reporting of enemy activity is something you'd want to do probably revolves around where you want the game to go. The broader your audience, the less 'sim' you're going to want and the fewer the delays the better.
If, on the other hand, you were aiming to 'conquer the niche' then increased realism that you can package up as gaming fun (no point doing anything if it's going to make the game less enjoyable) would be a better road to follow.
This has been a long winded post but, hey, you asked for ideas, so I thought I'd lay it out.
Cheers,
Lancer
The player seeing the enemy engaging 'unseen' enemy units is exactly how the helicopters work in the game at present.
The friendly AI wouldn't need to change (although any improvement would be welcome). The game and friendly A.I would work as it currently does with the only difference is that the player is issuing orders based on the map situation as they know it which may well be different to how one of their subordinate units sees things.
Here's an example that highlights the good and bad of the idea and the current implementation of how things work.
Say your platoon of APCs spots enemy armour ahead. They send that information up the chain of command to you, the player, at HQ. Because of the delay in pushing the information upwards and perhaps some interference from EW, the information arrives 15 seconds later (real time). In the meantime, you issue an order for the APC's to advance, unaware of the armour threat.
The map doesn't yet show the enemy armour because you haven't received the information. Your APC's know that the enemy armour is there in front of them but until their sit-rep update arrives at your HQ, nothing is shown on the map.
So now you've issued an order that is quite possibly suicidal for your APCs. What happens? Well there is an order delay before implementation and you may issue another order once your recieve the update and the map shows the armour. Either way there is going to be a measure of chaos as you APC's advance a bit then reverse.
The other possibility is that your APC's may have preset operational parameters (like those that can be set by the player and already exist) that give them a measure of self preservation, regardless of your orders. This opens up another discussion about what level of autonomy should your subordinate units possess? Should they do exactly as you tell them or be able to 'query' an order they might consider dangerous (and how would you code the limits of what's dangerous and not?)
Contrast the above with the game as it currently is. The player has godlike awareness of his own forces (no different) and your APC platoon is giving you instant updates of everything they see.
So the player looks at the map, see nothing in the vicinity of the APC platoon and issues an order for them to advance. Shortly after doing so the enemy armour appears in front of them. The player sees this instantly (no delay in the information moving up through chain of command, no radio drop outs, no EW).
What's changed? Your APC platoon will still advance into the face of the enemy armour because you, the player, has issued an order to do so. You can issue another order to countermand the advance and have them reverse but you'll end up with the same chaos as above.
The APC platoon could have the ability to act in self preservation in accordance with some player determined operational doctrine but that is identical to the above and runs into the same issues of how much autonomy you would want your subordinate units to have.
The big difference between how the game currently is (all enemy activity shown instantly) and one where the map incorporates information delays on enemy activity is that the player will be able to react to having issued a bad set of orders faster (no information delay) with the current setup than with one which incorporates a delay. The friendly AI would act identically in both situations and would have the same issues.
So the question, perhaps, is what benefit would incorporating information delays have?
It'd make things more realistic. Whether it'd make it more fun is another question but if you were aiming the game such that a military audience was part of the deal it would be, I'd imagine, a big selling point.
It'd give you some scope to play around with things like EW and terrain in terms of their affect on information delay. I can't vouch for EW but I have a lot of experience in using radio equipment in various terrains and even high spec military equipment runs into 'drop outs' and 'low signal' once blocking terrain comes into play (eg. a hilly map). Sending units into difficult terrain and you're going to have issues with communications.
One aspect that the game doesn't model (and the competition does) is the amount of radio traffic inbound and outbound from an HQ. The enemy is actively searching for the high signal locations so they can drop artillery on them and hopefully take out HQ's. What you could do here is give the player an option for the level of traffic they are willing to live with. At low levels the information works it's way up the chain as described above but at high level perhaps it would represent the player at HQ issues lots of 'queries' to his subordinates, eg. 'what's going on? / are those tanks still there?, etc.' and this could be represented by a global reduction in the speed in which information travels up the chain of command.
You could tweak this further by having a global modifier to information delays that reduce them when the player's HQ unit is stationary and increase them when it's mobile. So a player then has to weigh up the benefits of improved communications and reduced information delays vs. the risk of being identified by the enemy signals surveillance and being targeted by an artillery strike. There's some potential for interesting and fun decisions here with the player having to hop his HQ around interspersed with short periods of stationary, high comms, activity when he really needs to know what's going on. What you would be doing, in effect, is giving the player a measure of control over their command cycle with a risk/reward game mechanic.
Lots of stuff you could do here with information delays to, perhaps, nudge the game towards a more interesting and realistic direction.
The big downside would be that it would highlight any friendly A.I issues even more so than currently as the player would not be able to react as quickly as they currently can with instant, on-map, enemy activity on the map. As highlighted above, there would be no difference in the reliance that the player would have on the friendly A.I but information delays would serve to bring any deficiences into a sharper focus.
The friendly A.I is already an issue (I can live with it but a lot of other people find it a concern) that, I'd suggest, you're going to have to address in some manner with the next version. It's an issue not because it's bad but because of how you've implemented the issuing of orders (with a delay) that requires friendly units to act autonomously and in a perceived sensible and logical manner during the intervals when you can't tell them what to do.
Whether putting another delay into the game mechanics around the reporting of enemy activity is something you'd want to do probably revolves around where you want the game to go. The broader your audience, the less 'sim' you're going to want and the fewer the delays the better.
If, on the other hand, you were aiming to 'conquer the niche' then increased realism that you can package up as gaming fun (no point doing anything if it's going to make the game less enjoyable) would be a better road to follow.
This has been a long winded post but, hey, you asked for ideas, so I thought I'd lay it out.
Cheers,
Lancer
-
Elmo_Zumwalt
- Posts: 17
- Joined: Fri Aug 24, 2018 10:52 pm
RE: “Information delay” implementation
Wow, this topic has gotten a lot more traffic than I would have expected ... not that I’m complaining[:)].
Lancer,
I think the AI autonomy would need to be greater with an info delay, since the possibility of a situation such as yours developing and resolving before the info delay + command delay has elapsed is greater than with only command delay. So your hypothetical APC platoon might need to be able to evaluate the situation on its own and possibly call of the attack/movement in favor of holding in place or retreating. Such behavior is found in Command Ops 2.
Another factor to consider is lateral information sharing (that is, info sharing between units at the same level in the hierarchy).
Lancer,
I think the AI autonomy would need to be greater with an info delay, since the possibility of a situation such as yours developing and resolving before the info delay + command delay has elapsed is greater than with only command delay. So your hypothetical APC platoon might need to be able to evaluate the situation on its own and possibly call of the attack/movement in favor of holding in place or retreating. Such behavior is found in Command Ops 2.
Another factor to consider is lateral information sharing (that is, info sharing between units at the same level in the hierarchy).