Jump to content

Idea about creating a "simulated dyn campaign" - a


bflagg

Recommended Posts

Yeah.. The mistake appears to be using CPU power for everthing that is occuring on the map.. reguardeless if it's 200mi away or 5 miles away along with using DirectX8 instead of Directx9.

 

I am no game programmer and wouldn't surmise that I can talk on a even base with the

coders at ED.. but if I understand it right Dx9 is more efficient than 8, whereas fog, water, and haze would be rendered with better efficiency. (so why 8? besides install base) (this subject has been hashed thru in the past)

 

It also appears that they can't change it (to far into it) but only tweak it.

(I really do hope I'm wrong.. but the last 2 years falls right in line in what I'm saying)

 

I think for them to succeed is to do something similar to what FO is doing, mixed with

a solid DC and they will be around for quite sometime in the FlightSim world.

Otherwise they will begin to blur into the background with all the others..

Thanks,

Brett

Link to comment
Share on other sites

  • 8 months later...
It's not that nice. Flanker 2.0 and 2.5 had exactly such branching campaigns, and it was almost a disaster.

 

The problem is fundamentally one of mathematics. Every single level that the user progresses through such a campaign requires the campaign creator to create a geometrically increasing number of missions for all the different "branches," the majority of which any individual player would not even see. The amount of wasted work spent creating the "dead branch" missions rapidly overtakes any time saved by not programming a fully dynamic campaign. The campaign file size also ends up being huge and slowing down the user interface. And the result is that the campaign ends up being necessarily much shorter, reducing the play value for the user. "What? Six missions and the campaign is done?"

 

Tried it once, won't do it again. DC or bust.

 

Sorry to bring this up again, but in such a branching campaign, can the campaign creator simply not just string along missions together that allows missions of the previous "level" to be played? Like:

 

Mission A:

Success, go to Mission B. Failure, go to Mission C.

 

Mission B:

Success, go to Mission C (i.e. loop back up to the "Failure" mission of Mission A). Failure, go to Mission D.

 

And so on. The result of the campaign, success, failure or neutral, can be determined what "path" that missions take, which could simply be determined by the result of key missions. For example, if the player keeps failing, then it would result in a large, probably impossibly difficult mission whose result would determine if the campaign was an utter failure, a stalemate or even if the campaign is to be continued. The same mission could also be used in a "success" path as well, except that even though chances of succeeding are still the same, the mission itself wouldn't have that much weight on it on the campaign outcome.

 

In this manner, a lot more missions can be used and played, no? And not as many would have to be built.

 

I actually think a branching campaign is *much* better than a linear campaign. It just depends on how it's structured. Replay value can also be increased if the result of one mission leads to a random selection of the next mission from a "group" of missions linked with the result. Like, a successful Mission A would lead to a random selection of Missions C, D or E, while failure in Mission A lead to a choice between Mission B or F.

 

And then the results of those missions can be looped back up to play missions that were not chosen in the random selection.

 

Sure, some missions would still not be played unless the player deliberately chooses to fail (or something), but I don't think a DC is necessary.

sigzk5.jpg
Link to comment
Share on other sites

The amount of wasted work spent creating the "dead branch" missions rapidly overtakes any time saved by not programming a fully dynamic campaign.

 

Agreed!

 

open it up to 3rd party terrain and give us modders an outlet to turn our constant begging for changes and extensions into something productive.

 

Yep!

Dave "Hawg11" St. Jean

Link to comment
Share on other sites

I'm surprised that so many still support an enhanced mission editor - despite

the fact that every patch released makes missions 'incompatible' (more or

less). The BSO 'desaster' is the best example:

http://forum.lockon.ru/showthread.php?t=14488

 

All the fancy graphics, zillions of rendered polygons per second cannot hide

the lack of gameplay in this sim. I don't have the time to spend hours and

hours designing missions, but i have the hardware to run a mission generator

or dynamic campaign or whatever. Falcon has it's DC (real time), IL2 a

campaign (mission based) - they have their drawbacks, but they exist.

Oh, i still like Lockon, but i play it online exclusively - the offline environment

is too sterile and 'lifeless'.

 

But the series will end with 1.2 and all after that is nebulous.

Link to comment
Share on other sites

I'm using F4AF at the moment and finding the latest patch very good for the sim. Patches should always be compatible with missions if you are using a DC. Most missions in lockon still run but the tracks are most effected, some have to be remade because of improvements. As it is a single large mission can be created in FC that works in real time, such missions can go upto a month. FC has this capability which is quite unique but to generate an ATO you must first create the ability for the campaign to select the targets. The ATO then gives the target as a mission. From here if FC allowed the player to create their own flight plan and also displayed known threats and inteligence in the area then it could work. As it is now every single flights route appears on the map when you go online so its difficult to tell which one is your aircrafts. Also each aircraft has its own flight plan in MP, if you could create flight plans that allowed MP to work with 4 ship human flights then it would be much better. IMO bubble method is the way to go, branch missions are nowhere as good as a proper DC. One thing about F4s DC is success is based on the players performance.

[sIGPIC]2011subsRADM.jpg

[/sIGPIC]

Link to comment
Share on other sites

Sorry to bring this up again, but in such a branching campaign, can the campaign creator simply not just string along missions together that allows missions of the previous "level" to be played? Like:

 

Mission A:

Success, go to Mission B. Failure, go to Mission C.

 

Mission B:

Success, go to Mission C (i.e. loop back up to the "Failure" mission of Mission A). Failure, go to Mission D.

 

I actually think a branching campaign is *much* better than a linear campaign. It just depends on how it's structured. Replay value can also be increased if the result of one mission leads to a random selection of the next mission from a "group" of missions linked with the result. Like, a successful Mission A would lead to a random selection of Missions C, D or E, while failure in Mission A lead to a choice between Mission B or F.

 

It's not a viable option. You should do some maths to understand why. Branching means an exponential growth of the tree. A tree which must me manually compiled, with each mission being human-scripted.

You want a 10 mission campaign (quite small), with just a fail/succeed result? well, you need to build about 2^10 missions.

Actually the real formula is (I correct it from my previous post) ;) :

 

(b^n) - 1

 

where n is the number of missions for the campaing, i.e. the depth of the tree, and b is the number of possible outcomes of a mission (typically 2, failure/success), i.e. the branching factor.

 

So for a 10 mission campaign you should build 1023 missions ;)

If you plan to use a success/neutral/fail scheme, for a 10 mission campaign you should build 59048 missions ;)

you should see that it's quite impossible

Link to comment
Share on other sites

What D-Scythe was getting at was the approach to reuse missions at various parts of the campaign tree structure. Obviously this would decrease the required mission building workload. It's also possible to tweak only a few parameters of missions here or there depending on where in the campaign tree you are (like, less opposition on success branch compared to failure branch, but otherwise the same missions).

 

However the "logistical overhead" is still a nightmare. Starting at campaign step 4 or 5 you start to lose track of which mission file belongs to which branch, even if you don't have a complete binary mission tree for each step. Trust me, I tried it - and gave up for a reason. I'm with SwingKid here: the additional workload is heavy, the benefit not that satisfactory in the end (even though it's still better than completely linear campaigns, no question here). And, as has been pointed out, you run into compatibility problems with the next patch.

 

Should I ever have time again to dig into the mission compiling/decompiling area again, I'll rather invest in other features (in fact I'm full of ideas, but completely out of time ;)).

Caretaker

 

ED Beta Test Team

Link to comment
Share on other sites

In a DC there are only about 20 mission types and its the ATO that generates which ones take priority. The one thing that is required is a automatic flight planner. This generates an automatic flight plan with waypoints, IP, TGT, land, takeoff. With an A-10s DC it would simply be On-Call CAS, BAI, FAC , S&D or preplanned CAS etc. Then the targets that the ATO delegates to the A-10 squadron start with the closest targets working outwards with the exception of attacking logistics/supply convoys. Where you might fly further behind enemy lines and the result might be that enemy T80s run out of ammunition or fuel in that area. Could even go as far as having the AI go through a process of resupply on the battlefield.

[sIGPIC]2011subsRADM.jpg

[/sIGPIC]

Link to comment
Share on other sites

No, the 'one thing that is required' isn't an automatic flight planner. That would be one of the small things required. What about the rest of the campaign? ;)

[sIGPIC][/sIGPIC]

Reminder: SAM = Speed Bump :D

I used to play flight sims like you, but then I took a slammer to the knee - Yoda

Link to comment
Share on other sites

No, the 'one thing that is required' isn't an automatic flight planner. That would be one of the small things required. What about the rest of the campaign? ;)

 

I'll suggest you ask ED; they had a programmer coding a DC for LockOn.

They should have the concept in the bottom drawer, or not?

Link to comment
Share on other sites

Guest ThomasDWeiss

Even if a dynamic campaign was created it would be useless with LO as it is.

 

Lock On is a CPU dependent sim, that means that there is a finite number of units that it can handle.

 

Some units are resource hogs (one of the worst - the Su-25T, artillery is a killer) there are some AI actions that can crash LO and the terrain is itself quite demanding for the CPU to run.

 

Mission building is time consuming: to build one takes anything from a week to ten days. A complex dynamic campaign? months and months. We need a new Mission building tool, we don't have one.

 

We don't have random events.

 

We don't have AI triggers.

 

There is no way to exit an unit from a scenario.

 

A dynamic campaign would have to be a very simple affair, and as much as we could want one we would still not be happy with it.

 

What we could hope for is a Jane's style campaign.

 

And if we had a Jane's style campaign what would we get? to create a 10 mission campaign that would not be boring ( needing 30 to 40 missions to have some 3 to 4 branches ) a dedicated mission builder to create it would have to work on it for about eight months.

 

And for such an amount of work, we would need to pay to get it - and as far as I know, ED is not supporting yet a Falcon style model of business were an add-on is sold like Aeyes does.

Link to comment
Share on other sites

What we could hope for is a Jane's style campaign.

 

There was another Jane's title that shared some limitations with Lock On and still had a superb campaign system - Longbow II. Lower overall unit number (compared to other DC systems like Falcon4 or even EECH), only max. 16 flying units per mission (normally less), ground units mostly static or only advancing a short distance per mission, and no real-time campaign progress. Of course that system worked better for a chopper sim than it would for a plane sim, but then again we're getting a chopper with 1.2 :p

 

The various AI issues are still a problem, no question here, but it would still be entirely possible for the current Lock On engine to support a similar system. Certainly nothing Falcon4-like, but within that scope, it should work well.

Caretaker

 

ED Beta Test Team

Link to comment
Share on other sites

Guest ThomasDWeiss

Take this mission Caretaker

 

there are some 80 units in it

Map1.jpg

 

a closer look

Map2.jpg

 

it is a 140k mission - from my experience this is the maximum size for a mission

 

It uses a quarter or less of the map we have.

 

In terms of work - it took me a week to get it right.

 

Lock On has zero tools for the mission builder other than placing units, you have to fly over and over again a mission - so it is time consuming, there was a bug on this mission that CTD it every time, I had to find what it was, more time spent to debug.

 

I made this mission more static so there are more ways open on how to achieve the goals - you can't have too much dynamic action in a mission or it becomes impossible to fly it.

 

So what is my point ? Lock On does not need NOW a dynamic campaign - it needs FPS and CPU optimizations, LOD tweaks, better tools to help build a mission, RANDOM timing of events, RANDOM placing of a given unit.

 

It is of no use to have a huge map when you have to spread units so thin that it becomes simply useless going to and fro.

 

And on top of that - you have to build a mission that survives patches, who wants to spend a lot of time building a mission that will be unplayable in six months?

Link to comment
Share on other sites

What D-Scythe was getting at was the approach to reuse missions at various parts of the campaign tree structure. Obviously this would decrease the required mission building workload. It's also possible to tweak only a few parameters of missions here or there depending on where in the campaign tree you are (like, less opposition on success branch compared to failure branch, but otherwise the same missions).

 

However the "logistical overhead" is still a nightmare. Starting at campaign step 4 or 5 you start to lose track of which mission file belongs to which branch, even if you don't have a complete binary mission tree for each step. Trust me, I tried it - and gave up for a reason. I'm with SwingKid here: the additional workload is heavy, the benefit not that satisfactory in the end (even though it's still better than completely linear campaigns, no question here). And, as has been pointed out, you run into compatibility problems with the next patch.

 

Should I ever have time again to dig into the mission compiling/decompiling area again, I'll rather invest in other features (in fact I'm full of ideas, but completely out of time ;)).

 

Yeah. As I understand it though, Jane's F/A-18's campaign system was still that of a branching campaign (even though it was "semi-dynamic"), albeit there are a lot more factors and triggers available to determine which mission is played next.

sigzk5.jpg
Link to comment
Share on other sites

Thomas,

 

well this is a somewhat abstract discussion ;)

 

If you'd ask me whether I think a dynamic campaign is be the most important possible improvement for Lock On's gameplay right now, I would clearly say no. Better control over AI units and their sometimes not-so-intelligent behaviour would be far more important. As you're saying, more tools for mission building, and no question that performance improvements would also always be welcome.

 

Nevertheless I think a dynamic campaign would be possible, within certain limits, already now. A branching campaign in turn is not really the solution to the current system's shortcomings - especially with the version compatibility problems.

Caretaker

 

ED Beta Test Team

Link to comment
Share on other sites

Yeah. As I understand it though, Jane's F/A-18's campaign system was still that of a branching campaign (even though it was "semi-dynamic"), albeit there are a lot more factors and triggers available to determine which mission is played next.

 

Indeed, and with these triggers and other options it was possible to change the way the same mission played out each time it was flown. Branches were mostly used for different paths after a major campaign milestone was fulfilled or missed. I think the outcome of several previous missions could also be taken into account at these points, to determine which way the campaign would take next.

 

Creating mission dynamics with branches alone is quite tiresome though. As I said, I'm not against it, but I really doubt it's a major improvement in the end.

Caretaker

 

ED Beta Test Team

Link to comment
Share on other sites

Guest ThomasDWeiss

Probably if someone prepared an engine it would work - but I am sure we would end up with random units popping up here and there aimlessly wandering around until one bumps into one and then into another. One of the most difficult things is having an aim, a goal - otherwise it is just a bunch of units strewn around a map, that is the aim of a mission: to have a purpose.

 

With the means at hand - and the limitations - how to focus a campaign ?

 

That is how I approach a mission, find a purpose for it, for this, you would need to narrow a lot the scenario.

 

I have the impression what people want is not a dynamic campaign, but a random mission maker so they can go flying over a place where lots of thing go BANG BUUM!

 

I am quite curious to see what (if any) improvements 1.2 will bring, but I doubt we will see any changes, different AI behavior yes, but no new stuff.

Link to comment
Share on other sites

Lockon will never have a dynamic campaign, at least until the game engine is rewritten:

 

http://forum.lockon.ru/showthread.php?t=6595

 

surely a dynamic campaign could be played also with a dozen units (if they are dynamic) BUT a dynamic campaign is fashionable if it has a lot of units moving... well lockon simply can't handle this, even without a campaign engine, i.e. with units following scripted orders.

 

Anyway, the branch-tree algorithm for a campaign is useless because it's inefficient not by a constant, but by an order of magnitude... even if you cut the tree by some constant (i.e. if you find a way to script only 1/4th of the missions) you'd still have an exponential growth of the tree -> number of missions. So it's not a viable option.

Link to comment
Share on other sites

I have the impression what people want is not a dynamic campaign, but a random mission maker so they can go flying over a place where lots of thing go BANG BUUM!

 

Yup, that's quite often my impression as well... not that this approach doesn't have a charm of it's own ;) Even a very basic random mission generator has the advantage of unpredictable encounters though. That's also an important factor, even though I prefer missions with a "purpose" myself.

 

 

Lockon will never have a dynamic campaign, at least until the game engine is rewritten

 

Only if Falcon4 is your only idea of a dynamic campaign. As I said, there are alternatives :)

Caretaker

 

ED Beta Test Team

Link to comment
Share on other sites

Anyway, the branch-tree algorithm for a campaign is useless because it's inefficient not by a constant, but by an order of magnitude... even if you cut the tree by some constant (i.e. if you find a way to script only 1/4th of the missions) you'd still have an exponential growth of the tree -> number of missions. So it's not a viable option.

 

It's inefficient, but no, you would not necessarily have an exponenetial growth of a tree. Just take a look at Jane's F/A-18's campaigns - there are probably only 12-18 missions for 6. In comparison, if a true branching campaign was implemented, that number would be at least 60 for a 6 mission long campaign.

 

Such a system obviously isn't useless either if it was such a huge success with Jane's F/A-18. Although to be fair, that is the best such system to date.

sigzk5.jpg
Link to comment
Share on other sites

Guest ThomasDWeiss

Never having flown F4 I can't comment on it, but TAW had Wargen and it had a very little variety in missions - and Jane's had a great drawback, I tended to try to win every mission, so I ended up flying the same missions and never saw the alternative ones.

Link to comment
Share on other sites

As I said, there are alternatives :)

 

Hm, and what are the chances to see anything become implemented?

And in which sim? 1.2? Tank killers? F-16 sim in 2010?

Geeze, you can discuss in thin air all night long, but you can't

give any reasonable perspective - not your fault, of course.

 

There won't be a DC (like Falcon) for Lomac, the engine can't handle it.

I can't imagine a mission based DC, the next patch will break it.

Will there be enhancements to the mission editor? Naah, 1.2 will be

released soon (within a few months), so ED can tell us about it.

Or 'Tank killers', the sequel, might have all these great enhancements

and new features ... but i will believe it when i can play it ...

as Stormin said: No promises ;)

Link to comment
Share on other sites

The Bubble method must be used as large terrain area will be the minimum standard in the future. Also much larger battles can be fought using such a system rather than bogging the cpu down with objects that are thousands of kilometers away. Automatic flight planner and ATO are required if you wish to create a DC like Falcon. Branching campaigns are not DC they are just missions tied together with a yes/no switch. You only use that method for campaigns that involve actors, they are no good for a DC.:cool: Think of it this way 10,000 objects vs 1,000,000 or 1,000,000 players.

[sIGPIC]2011subsRADM.jpg

[/sIGPIC]

Link to comment
Share on other sites

Hm, and what are the chances to see anything become implemented?

 

Oh, right now, they are still very slim :D As I said, this is more of an abstract discussion.

 

There won't be a DC (like Falcon) for Lomac, the engine can't handle it.

I can't imagine a mission based DC, the next patch will break it.

 

Any "mission-based" DC would have to include automatic mission generation, which is far from trivial, but still possible already now. This would then be relatively easy to adjust to a new version, unlike the mission files themselves.

 

The Falcon4 campaign has two additional features: the bubble system and the real-time approach. The former is a nightmare to get right, so I personally won't expect this for a while still in any Lock On successor; the latter is a luxury that isn't quite as important in my opinion, but it's also easier to include from a development perspective.

 

but i will believe it when i can play it ...

as Stormin said: No promises ;)

 

That is certainly a reasonable approach... ;)

Caretaker

 

ED Beta Test Team

Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...