S! Tailgunner!

Keep in mind that dial-up users DO have enough bandwidth for online play. We had massive online servers in RB3D when MOST users had dial-up. I sat down and figured out exactly what the requirements were once - something around 12 KBPS on the client and 12 KBPS per person on the server. I've done similar tests with FB and the numbers are consistent. That doesn't mean every game will have similar requirements, but certainly it is possible to write a flight sim that plays well over dial-up.

I'd also keep in mind that any new sim built with ideas from this thread would be 2 years away at a bare minimum, so the online environment won't be today's environment but rather the one 2 years down the road. That means a greater percentage of broadband/always on users and a greater percentage of dedicated phone lines for the remaining dial-up users.

And I don't think the assumption is that everyone will want to be connected to play. I think the assumption is that if connecting is a requirement then it will be one that the vast majority can live with. There may be a number of users who would prefer not to connect, but I don't think there would be many who would be unwilling to connect. Heck - there might be users who would prefer to play with their computers turned off. You can't make everyone happy...

Developers need to innovate, and much of the innovation is going to be based on interoperability over the Internet. As more and more users get connections that are always on, there should be a trend toward using these connections to enhance the gaming experience. I don't think that someone should say 'you can't do that because that's not how it is done.' I think there needs to be a better reason not to do something than just that it hasn't been tried before.

With regards to some being dial-up as Tailgunner says, yes as I mentioned and as Wall-dog says this game would not be out today but using future Internet connections (i.e. more always-on users).

Also if someone wants to be "offline" and wanted to completely opt out of any online content, they get the dodgy AI and the scripted comments/missions, everyone else who wants the "enhanced" game chooses an online connection. easy

The new Novalogic game "Joint Operations" hosts 150+ player servers with tanks, helicopters, jeeps etc full of individual people all moving around doing their own thing, bullets flying etc all handled superbly considering!

*oops*

...goes to blackboard...in front of whole class too...

tailgunner
Tailgunner
tAilgunner
taIlgunner
taiLgunner
tailGunner
tailgUnner
tailguNner
tailgunNer
tailgunnEr
tailgunneR
TAILGUNNER

now, this one mistake won't happen again.

No I never thought of the "tailwind" thing until you mentioned "insult"--if that is what you are thinking. Pure mistake. Sorry.

McGonigle, are you aware that the most interesting (IMO) online virtual arena for World War I aerial combat has already been conceptualized:

Birds of Prey virtual arena (Part 1)

Birds of Prey virtual arena (Part 2)

None of us can know if BoP will ever see the light of day, but its design concept is still as valid and captivating today as it was years ago.

Finished reading here. I stopped at Tailgunner's note to learn how to pay attention to names (and I hope its clear I did not just cut~n~paste Tail's [full] name but typed it until I can now have it in long term memory).

For the simhq board, one example of AI programming that would be useful in both offline and online flight simming (dull dull description followed by explanation in next post below)

The tactical space game O-R-B (Orbital Resource Base) allows the player to command hundreds of AI ships either individually or collected into groups. The ships conduct their own very simple combat. The player may select what formation the groups of ships will fly in. For example "X" formation, line formation, Vee formation. The corresponding idea here is to allow flight sim mission builders to define formation that AI planes fly instead of being limited to a single default type (thinking of Forgotten Battles here).

The O-R-B player may take any number of individual ships and order them to group into a single fleet. In O-R-B, if this number exceeds 20 ships then the ships form multiple formations of 20 ships (or whatever number is left over for the remaining formation). These multiple formations are easily controlled and act as a single entity (fleet) but still retain individual ship behavior but confined within its formation.

Even better, the player may define AI behavior for each ship or groups of ships. The available behavior includes total agressive and attack anything within range, shadow or escort another ship or group of ships, total evasive behavior, or "neutral" behavior where the ship/group ignores any targets or threats to itself. Further behaviors include maximum distance before retreating for re~arming weapons. Many other behaviors that I did not explore in my short time playing the space sim.
~ http://www.o-r-b.com/
-----------------------------------------

Sorry for the long description; the important "flight sim" points for O-R-B AI are...

(1)Flight sim mission builders defining number and formation of aircraft from 1 plane up to hundreds of planes.

(2) Flight sim mission builders defining the behavior of AI aircraft. We give an example of hidden AI behavior not documented in Forgotten Battles manual or readme files. If a single engine AI bomber follows a "normal fly" waypoint, it will drop bombs and attack any enemy aircraft within standard AI engagement range--no matter if the enemy plane is not a threat--even a lowly Fi~153 Storch. However if the AI bomber is following a "ground attack" waypoint then it will ignore any aircraft threats to itself and keep bombs and fly to surface target.

The mission builder should be able to define which of these two types of behavior the AI bomber will follow--with the AI programmers possibly coding a random chance of AI ignoring ordered behavior as in real life ~~> THIS last point is a perfect example of how one may "improve" AI programming.

------------

As Wall Dog points out above, online simming could benefit hugely from this kind of "improved" or "new" AI programming--improved and new for flight sims. But why do we see a contradiction in WallDog who states AI are not needed because offline simming is not needed, but then WallDog posts that AI are needed to "help" online flight simming (which I agree with).

WallDog:: Understood! In this thread we are not talking the same flight sim language. You are talking IT Professional language alone. By "improved" AI I mean several things including but also far beyond the "dogfight" AI that captures the entire imagination of some here. walldog?


WallDog on AI Chess:: WallDog you know very well that Chess AI programming checks all possible future moves for as many future game turns possible given a chosen time limit for the computer, and the number of future moves looked at determines the computer horsepower required. I had a Chess game written in Fortran on my old Pentium 133MHz computer and it ran great and was very challenging although it was not the world class champion software of the type developed by IBM for Gary Kasporov that you try to compare with a home computer.

Worse, I don't think flight sim AI is comparable in algorithm to Chess software (simhq Debate?).


McGonigle:: Yes. WallDog claims to be an "IT professional" and his/her ideas are related to that area of networked computing, not that there is anything wrong with that, but why the desire of "IT Pro" to eliminate private flight simming in one's home?


Today, Online AI aircraft use the same AI programming as offline AI programming (a point missed by WallDog). Online war missions will be created with the same tools as offline missions are created. For WallDog's IT Network flight siming, we need improved AI such as...

Original DOS Flanker 1.0 jet sim of 1994 had heat seeking missiles that were often distracted by the sun--imagine an old 1994 flight sim created for the 486 cpu that could model AI missiles blinded by the sun. Similarly, flight sim AI pilots' vision being effected by the sun will allow humanoid players who are plugged into WallDog's IT Network to bounce the AI from out of the sun--a brutally crucial needed AI gameplay element for prop sims and especially for World War 1.

I call this "improved" AI programming because Forgotten Battles does not model AI blinded by the sun. Programming this feature would be an improvement for that sim at least. Do the other recent flight sims model AI blinded by the sun? What I DON'T understand is why the needed improvements for AI that would help create popular demand for ever improving IT Network flight simming is being denied by WallDog.


(*sigh*... more )

WallDog caught practicing Deception @ simhq.com:: Most flight sims define themselves as being online and offline. An IT-Pro should know about And/Or.

Also, we are talking about flight sims here not "programs" in general. Be very careful simhq.com. WallDog IT-Pro just betrayed a desire to make ALL computer programs IT Network only.

Why are you trying to confuse people here? This does sound like a Microsoft "IT professional" spokesperson. Very Confusing.


Online play is great
Offline play is great
WallDog IT-Pro will not allow flight simming unless you are plugged into WallDog's IT Network (and presumably paying for the New Privilege of computer flight simming).

Not so confusing after all. :p

I think it is pretty clear that AI is an essential for any flightsim. So... I ran some tests.

I have a boxfulof different sims to choose from, so I gave them all a spin to test out the way AI reacted.

I gave them a good turn fighter, and they all tried to stick on my tail. Some skill was shown by IL2 AI. Novices tended to fire earlier. Experts tended to hold fire till they were close. Most seem to give 'Ace' pilots a LOT more 'G' tolerance. CGS3 is a real naughty boy here. It's difficulty levelling is a bodge of the worst sort. Give them an energy fighter, like, say a P47, and they don't have a clue!

Not fair..I know. I knew they would do this!

So, I switched to other things! Can the AI find you in thick cloud.... surprising how many AI's can do this without radar....Night flights are worse! All have excellent night vision.

In some cases, AI is clearly ignoring me until reach a certain 'range' before it evades.

The AI never flys the plane to the point of stall

The AI has a predictable set of manouvers to shake off enemy ( IL2 'dive and turn' ) which it uses regardless of situation

Now, I could go on, but I think this pretty much outlines why people are not happy with current levels of AI. They are too scripted in their actions. They are too predictable in their reactions. They never surprise you after you have played for a while.

This is basically impacting on immersion. Immersion is King. I agree that blurring of Online /Offline may be a way to go, but the AI needs to improve in some key ways.

1) It must share the limitations of the pilot in terms of stress and G tolerance.
2) It must share the problems of maintaining visual contact where appropriate.
3) Even with the limits imposed by swcripted AI, the AI pilot should have 3 or 4 options to choose from when reacting e.g. Turn left, dive and extend, split S, turn right with increasing options for 'harder' AI

There should also be AI options for the chaps on the ground! Flak, vehicles evading, that sort of thing. Shoot up a comvoy and they should do more than just stop coz the roads blocked!

Now... I know there are sims that do SOME of this...but none do ALL. Making AI that reacts like a human is still a long way off, but surely the next generation of sims should be looking at improved AI behaviour beforefancier grpahics ?

Oh...yes...and teh Best AI award from my test....


Red Baron 3D. They gave me the best run for the money. 2 seaters evade like anything, and always manouver to get their guns to bear. Progress... oh well

Lexx,

You've been making pot shots at everyone on here for two days now. Do you want to debate me or just throw adolescent insults?

You talk about programming better 'AI' and then when pressed about how to do it you say that the mission builders should script it. Now you want it both ways - you want the AI to control aircraft in some posts and you want people to control them in others. And you say I don't have a consistent vision here?

Yes - it is true that IBM's Kasparov Chess computer is good. It is as good as Kasparov. I think they split games the last time they played. But you take that a step further and say it isn't a fair comparision. Come on now! Flight combat is exponentially more complex than a chess game. How can it NOT be a fair comparison. Because Kasparov is good? He's still human. We have people in our community that are awefully good at flight sims. Can you make AI better than it is? Yes, you can. But can you make it truly challenging? It sounds like you've given up that part of your argument though. Now you just want it scripted.

I can find every bad debate tactic in your posts. Have you heard of 'Ad Hominum' techniques? That's when you can't attack your opponent's position so you attack your opponent instead. You say things like 'But why do we see a contradiction in WallDog...' or ' Yes. WallDog claims to be an "IT professional" and his/her ideas are related to that area of networked computing, not that there is anything wrong with that, but why the desire of "IT Pro" to eliminate private flight simming in one's home?' or better yet 'Today, Online AI aircraft use the same AI programming as offline AI programming (a point missed by WallDog).' That's just a handful of them. There are more. And every single one of them intentionally misrepresents my position. You aren't even arguing with me. First you misquote me to make me sound like I'm saying something other than what I am, and then you argue with that misquote.

Come on Lexx. Do you want to debate, or just make childish attacks? The rest of us are trying to have a rational discussion about the future of the flight simulation community. All you are doing is taking pot-shots at people.

Just want to chip to say how much I agree with the this:

"give the player the chance to ENJOY THE GAME THE WAY THEY WANT TO ENJOY IT."

Absolutely. Switchable options, then everyone is happy. (Unless it complicates online but I'm not getting into that...)

McGonigle, your pilot-centred idea is very close to my idea of the ideal flight sim too. I suppose what I want is not so much a flight sim as a pilot sim. I'm more interested in the experience of being a combat pilot than I am in the performance statistics of a bunch of aircraft. And current flight sims always intrude the interface between me and the air combat experience. Sometimes it's a good, immersive interface (EAW, RB), sometimes it's a constant reminder that I'm just sitting at a computer manipulating a program (Il2). But it's always there, in the way.

Finally a question for anyone who knows about such things - how does flight sim AI work? Is the AI programmed with a set of preplanned manoeuvres, or does it make inputs to the controls on the fly, moment to moment? Does it aim to point the nose at a selected enemy, or to place itself in a given point in the sky? How is the AI told about the capabilities and limitations of the aircraft it is flying (eg don't try to climb below certain speed etc).

My own experience of programming AI (if you can call it that) is extremely limited (for a very simple Age of Sail game written for my own enyoyment), but one thing I found is that very simple rules (maintain x range from enemy, keep presenting broadside where possible) can produce suprisingly complex and believable behaviour when set going. But how to do 3D movement and all the complexities of flight would defeat me completely.

S! Von Stalhein!

Flight sim AI has been programmed both ways in different sims, with varying degrees of success. Really it is up to the developer how they do it.

A rules-based AI has several strengths. If you program AI for different aircraft seperately you will see the AI stay within the basic flight dynamics of their aircraft. You can for example program such that AI pilots in a certain plane will stay above a certain airspeed or climb whenever a certain event occurs or whatever. YOu can make pilots basic 'energy fighters' or 'angles fighters' based on aircraft. The problem with this type of programming is that it is very predictable and does not really lead to complex maneuvers.

A sim can also be programmed to use pre-set maneuvers. The obvious strength here is that you can then use more complex maneuvers. The weakness is that those maneuvers may or may not be appropriate when they are used.

Most flight sims today use a mix-and-match of both techniques. They will have pre-set maneuvers that the planes follow, but select those maneuvers based on a rules-based criteria. This leads to both the best and worst of both worlds, where the planes use predictable maneuvers for different situations.

And therin lies the rub.. You have to define rules to make the planes 'smart,' but doing so makes the planes predictable. You can add a level of randomness to the maneuver selection, but then you go right back to having planes use maneuvers that are not necessarily appropriate. The more random the planes are, the less appropriate their maneuvers will be. The more appropriate they act, the less predictable they become.

The best AI can be led-along like a dog on a leash. Good human pilots over time learn how the computer reacts to given stimuli, and then use that stimuli to set the computer up.

There are just too many variables. Altitude is a variable. Speed is a variable. Plane types are variables. Relative speed and altitude are variables. Direction is a variable. Distance is a variable. The number of planes in a fight all present new ranges of variables. Clouds create variables. Visibility is a variable. Ammunition loadouts are variables. How do you program around all of these variables (and I only named a few) to choose the right maneuver at the right time? People do this in a fluid, intuitive way that computers are not capable of. People are not computers. People do not just 'process data.' People actually think.

In chess, do you choose your move in a fluid environment? No - you look at each alternative and at the consequences of each action. There are a finite number of moves, and even when thinking several moves ahead there are a finite number of strategies with a finite number of outcomes. Chess is the perfect environment to show-off the 'capabilities' of AI. Chess is complex enough to make AI impressive, but not complex enough to make it impossible. Think about how really good chess players play - they almost emulate computers! What a great fit for computer AI! Flight simulations are a completely different animal though.

Since the inception of flight simulations we have been asking for better AI. Over time we have gotten better AI, but it still isn't good enough.

Here is a quote from Tailgunner in which he is saying what AI should do:

Tailgunner wants the AI to choose from alternatives. This overlooks one critical fact - computers are incapable of making a choice. You can 'emulate' choice, but how do you do that? Either you make it random - in which case the 'selection' may not be appropriate, or you make it 'rules based,' in which case it will be predictable.

That is not to say that you should't try to make 'better' AI, but it is to say that we should recognize that no matter how 'good' we make it, AI will never be 'good enough.'

So what can we do? Well, we can keep trying to 'get blood from a rock' and watch our genre shrink while other genres grow, or we can try to program things better, focusing more on experiences and less on pre-concieved notions of what constitutes 'online' or 'offline.' If we can break-down those traditional barriers we can start to experiment with environments that are neither 'online' nor 'offline,' but really a mix between the two. And then we can start to do some really intersting things like use real people when possible for the 'AI' portion of an offline campaign.

Finally - Lexx did make one point I want to address. Lexx said that most simulations today are both 'online' AND 'offline.' Come on Lexx... There is a difference between an environment in which a simulation can do both at different times, and one in which the simulation is using elements from both at ALL times. What bothers me about your argument Lexx is that you knew that difference when you made your post.