Lock-On's limitations and the "Bubble" system

[Dmut]

Fly a campaign in Falcon. Hit the outside key in Falcon and cycle through units. You won't cycle through the 1000-plus-units which are running in the campaign. You will only cycle through the units that are loaded within the player's bubble. :)

 

Now do the same in Lo-Mac. Eventually, you will cycle through every unit that's been loaded in the mission. ;)

hmmm, you call that an advantage when you can't see\controll all units in scenario? I call it drawback.

 

yes, I perfectly understand that F4 uses a "bubble" of zero size when user is out of mission, but it's doesn't prove that it is better.

[Dmut]

I respect your point of view, however I stay on my - from what I can see as developer, DC is not contain or somehow connected to "bubble" system, which used for simplification of units calculation when user is not in mission or too far away from specific unit.

 

But, if you want to shift gears and would like me to point out why I think the bubble method would be very useful in the sequel to Lock-On, I'd be happy to.

well please do. because I see only one positive side of using "bubble" - it simplify calcalulations and saving CPU time.

[ALDEGA]

DC is not contain or somehow connected to "bubble" system,

 

I see only one positive side of using "bubble" - it simplify calcalulations and saving CPU time.

In a DC you expect to have a medium to high number of AI units. This brings performance issues. Optimizations such as the "bubble" system can help to fix performance issues.

 

Lockon runs today on hardware that is almost 7 years newer than the hardware from F4's time. Yet it can't handle the number of AI that F4 can on today's hardware.

 

However I don't know how many of you tryed to program the AI, for those who did they know that during programing process of AI a "tree" should be elaborated.

Don't forget to mention there are big optimizations possible using systems such as alpha-beta pruning and minimax.

 

AI activity inside the "bubble" does not have to be "realistic". It occurs outside the player's view. The "simulation" outside the bubble could be as simple as a simple "unit A has a 70% chance of destroying unit B". (for units outside the bubble that do not interact with objects in the bubble)

[Dmut]

The successor to Lock-On's FM will only become more and more complex. The graphics will become more system resource intensive (as shown with the graphics upgrades of the SU-25T in FC). If they add a fully clickable pit, as ED's proposed, still more system resources will be drained. Something's got to be compromised, IMO.

I agree.

In a DC you expect to have a medium to high number of AI units. This brings performance issues. Optimizations such as the "bubble" system can help to fix performance issues.

I agree.

The "simulation" outside the bubble could be as simple as a simple "unit A has a 70% chance of destroying unit B".

and that is the reason why some (including me) don't like bubble system - the result of units performance depends on distance to player.

(for units outside the bubble that do not interact with objects in the bubble)

bubble would be a perfect way to simplify missions or campaigns - IF ONLY it(bubble) was static... since bubble is moving with player - then player can decide what's result of units activity he likes more - when specific unit is inside bubble or outside, and then move it's plane accordingly. with this method player can "affect" those units on "bubble" boards, which he doesn't ever see or somehow notice.

[ALDEGA]

I agree.

since bubble is moving with player - then player can decide what's result of units activity he likes more - when specific unit is inside bubble or outside, and then move it's plane accordingly. with this method player can "affect" those units on "bubble" boards, which he doesn't ever see or somehow notice.

The player can always decide where to fly and thus affect the course of "history" depending on his choice. Normally however, you are restricted to the mission waypoints that you were ordered to follow by your superiors and restricted to engaging the targets that are specified in the briefing, instead of flying off to a random place.

[eXor]

I agree.

I agree.

and that is the reason why some (including me) don't like bubble system - the result of units performance depends on distance to player.

 

:confused:

 

If Plane A is superior to Plane B resulting in Plane A beating Plane B 70% of the time, what difference does it make if you model this statistically instead of in real time? The result will be the same. The difference is that the former method is far more efficient and thus allows the game to cope with many more units.

 

If I'm flying along with Plane A and Plane B duking it out beyond my sensor and visual range, why should I care what they are doing? Their actions cannot affect me in any way.

[Dmut]

sorry guys, it's 3:30AM here, it was a working day with a lot of damn work, and I feel my head is going to spit my brain out of bubble :D

see you tomorrow.

[GGTharos]

Hawg, I think the bubble system is more complex than what you describe.

 

The bubble isn't only placed around the player, but around ground objects as well, with different 'visibility' parameters depending on what enters this bubble - eg. ground targets are resolved at shorter distances than air targets, or SAM sites (SAM sites do in fact have pretty large bubbles AFAIK)

 

But there are still issues with the bubble system which need to be corrected. I agree with you that it is A way to make LOMAC's successor better and it is NO DOUBT being considered.

 

I don't agree that its useful for LOMAC, or that it eevr will be because of the work required to make it work, IMHO - specifically on the part of the devs.

[GGTharos]

No, there are additional bubbles, because the AI's fighting each other need to resolve their fights. See the logic?

[GGTharos]

Quit quoting that.

 

THis isn't how the bubble works ;) Each object has a bubble and when an opposing object enters it they are resolved in order to do battle.

[GGTharos]

You need to read some more recent Falcon specs. Because quite frankly, you're wrong. Quit reading out of a maula and go ask the community.

 

What you're quoting says ZERO about how the bubble systems works. If youw ant the details, look elsewhere. (Frug's). THEN if they tell you otherwise, you can tell me I'm wrong.

[GGTharos]

Actually, you didn't prove anything. You quoted a bunch of irrelevant things which don't actually describe the workings of the bubble system. I don't have to do your homework for you. You do it yourself.

[D-Scythe]

Actually, GG, I think (in F4 anyway) there is only one bubble, and its around the player. Objects outside this bubble don't have their own bubbles.

 

However, the combats/engagements outside the player's bubble are decided in a manner like two units engaging in a strategy game, so it's not just as simple as probability tables. In F4, if you observed the flights in a campaign between AI, you'd see its perfectly normal for both sides to bug out - it's not always just a matter of meeting and cancelling out each other in a probability chart, and there are some AI/CPU cycles involved to determine who is killed or not killed.

 

But that point is moot - Lock On V1.2 is not going to introduce a bubble system, so I don't see why we are still talking about this. Why not discuss something that the devs *can* change?

 

Even in the fantasy case that a bubble system is implemented in V1.2, no dynamic campaign is going to work the way its supposed to with the AI as weak as it is in Lock On.

[GGTharos]

Yep, I just re-read RP5.

[Dmut]

ok, I'm back and ready to continue this loooong discussion.

Let's exactly define the term "bubble" we will use - I can define it as an area around player's plane, where all unit in the game perform individually, with all possible (to specific game) details of performance and complexity. Units outside to "bubble" performs with simplified logic and result of units action is based on probability tables.

Anybody want to add anything to this definition to "bubble" term?

Now I want to explain why:

a) I don't like calculations, based on probability tables.

b) I don't like two different kind of calculations in game.

 

About a) - let's say we have 1 Abrams and 1 RPG man. What's probability of 1 RPG can defeat 1 Abrams? from my point of view it's about 0% (0.01% if RPG will be able to make a hit by rocket directly inside a A1's burreel, when HEAT is loaded :) ) So we will have a 0.01% in a table.

ok. let's take 10 RPG men and 10 Abrams - I believe 10 RPG men will be able to destroy 1 Abrams in such situation, which point us to 10% of probality. Also we should keep in mind a kind of place where such engagement is taking place - plain terrain or urban envirounment - those factors will have a great influence on combatants probabilities.

it's not about "red" vs. "blue" flamewars, just a simple example.

Now I have next question to F4's adepts - what's set of factors bubble engine is including in calculations based on probability tables outside of bubble?

 

now about b) since bubble is moving with player, we theoretically can get a situation, when AI units leaving or entering bubble while making battle engagement. Simple example - let's say plane A engaging plane B inside a bubble and very close to bubble's borders. Plane A manages to get on B's six o'clock (because B engage a target C in that time). Plane A had launch AIM-120 to B few sec's ago and now B has an incoming and don't has enough time and energy to perform a 3-9 maneuver, means it will be dead in a few sec. But in this time player's bubble move forward and those two A and B getting outside a bubble. Then bubble engine takes a "A vs B" probability from table - 10% and bang-bang - plane A is dead. That's why I don't like 2 kinds of calculations.

 

ps: anybody know how many bubbles exists on scenario when multiplay taking a place?

pps: sorry for long post, typos and grammas.

[D-Scythe]

I think F4's campaign engine is a bit more advanced than that. As I said before, it's not just probability tables, but rather more like strategy game. It actually keeps track of the number and type of weapons on-board each AI plane as the mission progresses.

 

For example, a two-ship F-15 flight might engage an IL-76. When an F-15 fires an AIM-120, that missile is tracked by the campaign engine inside or outside the bubble - it's just that it's Pk is calculated differently depending on whether its inside or outside the player's bubble. Everything is tracked outside the bubble - up to and including cannon rounds - but just in much less detail and a larger scale. It's not a matter of simply have probability where the F-15s meet the Il-76 and simply 'cancels' it out because its threat value in the probability charts is greater.

[GGTharos]

Hehe, there may be a better way to 'judge' the outcome of such battles than straight statistics ... you could employ AI methods or at the least more complex (weighted) formulae for the calculations.

 

It would probably work relatively well, but it may require some extra processing power depending on how its done.