Contest Time

Ok very simply its been put to me by some of the members that we look into another contest. I'm all for it assuming we can gather enough interest.

The idea that has been suggested is...

Quote:

---

what about a cannon AI? Here is the scenario: there is a airplane/car/ship/spaceship/whatever moving on the playfield. The task is to code an AI for the gun so that it makes hits on the object. The challenge would be to determine what path the object is taking so that the gunfire would hit the object.

* the gun would stay still, but it could rotate
* the gun would get information about the angle where the object is and how far away it is
* the gun would know where the object is heading to (ie. the end position)
* the object could vary its speed and path, but would be time limited to get to its target
* the object would never die, but hits are counted
* the object would have a speed limit
* the gun would have a limited ammo
* the gun would have a reloading time
* the gun would know the speed of the bullet

---

Suggested by merri. If we can gather enough interest I'm willing to run it. So please vote away :)

And have any comments, then post here but please do not post chitchat keep this as clean as possible :)

Pino

Sorry, don't have time to participate in contests. :(

I guess the contest would end somewhere in late January, so even though the next few weeks are busy, there is plenty of time in January. Unless you're busy in January, too :)

What language?

Are you guys going to release an interface definition for the cannon?

Do the bullets have a blast radius or is it only direct hits?

For the blast radius, it is undecided, though I initially thought to have direct hits only. Currently the language is VB6 only, but I've made it in object orientated fashion so it shouldn't be too hard to convert the project to VB.NET as most of the logic should stay the same. Can't tell well as I don't do .NET myself.

There is some information in the future contest suggestions thread: http://www.vbforums.com/showthread.p...wpost&t=306604

I'm making the engine a bit too complex if we think about it only for the contest... it is more like a full game engine that is shaping up. Though that shouldn't bother you too much, it doesn't change the contest side at all even if there are things that aren't required by the contest.

I can't do anything AI related, I'll also say no.

Maybe Merri will be the only one competing (and end up in 2nd place anyways :afrog: )

I won't be participating since I'm making the engine ;)

No, my programing level is still too low for that, but if there was such a contest, i would sure like to get my hands on the winning code ;)

I'll enter if I can do it in Assembly :D

I'm definitly up for it as I have said many times in the contest suggetions thread. It sounds like an awesome challenge.

To those of you that say you aren't up to the challenge, I say you should really try it. I had no clue how to write a puzzle solver in the last contest, but I learned a TON while trying. If you think you can't do it now, then why not try and learn something as you go.

The only way this contest (or any contest) will be fun is if we get a lot of interest and participation.

We will open it to

vb6
Vb .Net (i'll convert merris engine with some help)

I thought I should mention that efficient algorithms for doing this already exist.

Well the same was true about the sudoku solver, but that was stilla lot of fun.

There was a bunch of activity in the contest suggestions thread in the last couple weeks and now that there isa cool idea ou there, everyone seems uninterested. Is anyone up for this?

It just doesnt sound like fun to me, sounds like math

You dislike programming?!?! A lot of it is math... dislike it and you shouln't be programming unless you are a masochist.

What did you think about the pathfinder contest? It seemed to be pretty popular and it is pretty close to this in the way that you are given just a simple interface, yet you need to figure out how to use that interface to the max to make it do something you want it to do. Believe me, figuring it out and making it work is a lot of fun and not "just" math.

So people seem not to be very interested about anything game programming related? Or just dislike me or my suggestion? Or both? Or all?

If you know the destination, and the time that it gets to there, then it wouldn't be much of a challenge. The obstacle is if it changes speed and how quickly the cannon operated. I said I might be interested. I'd have to look at the VB6 rewrite of the class, and throw around a few ideas.

Those that have done the same thing in the past would have a big advantage, I'd think.

Doesn't seem like it would require a whole lot of code, but I've never tried to do it, so I wouldn't swear to that.

There is gravity and air pressure to make it harder. It is isn't like shooting in space where bullets go directly to their target. So it is a sideview, not a god view.

Dglienna, You won't know exactly where the bullet is going. Your cannon will only receive limited info and only occasionally. Maybe the object went behing some clouds (on and off) for 10 seconds and you can still shoot him but you have to guess where it is when it is behing clouds (obviously without it traveling at a constant velocity). Thats where the fun comes in.

Finally a good use for the maths i learned. :) Is this going to be like a 3D game or is it 2D?

Just 2D. Coding a 3D engine would take all too much time. I don't have any copy-paste code for this either, so I don't know when I have the engine ready for the contest. Been long since I've coded anything game related (something that can't be counted just as doodling).

ah moon patrol, those were the days...

Sounds like fun, even though i have no experience at this whatsoever, and i doubt ill even make the hand-in date :)

So we have the possibility of 7 participants (counting the Maybes) plus a couple people from the contest suggestions thread (Devion, Raedwulf) so I think we might have enough interest to get this contest off its feet.

I'll give it 1 more days.

Then i'll let you know, and i'll have a chat with merri about the engine

Quote:

---

Originally Posted by Merri

I guess the contest would end somewhere in late January, so even though the next few weeks are busy, there is plenty of time in January. Unless you're busy in January, too :)

---

I got some exams in January but if you get it just during the christmas holidays I'll be fairly free :).

Sound's interesting, if I have time I'll be up for it.
just a thought; Instead of haveing a gun shoot at a target, it'd be way more fun to have each 'gun' be a bot and fight against each other in an arena.

I'll have a go, either in 6 or .NET, whatever I feel like at the time :wave:

-TPM-: To put it short, that is possible to make, it just requires a bit more coding by me. I did have a thought about doing the required stuff for that as the engine has come out pretty well and is quite flexible (codingwise that is).

I don't mind if people wish to change it to a botmatch instead, though it probably means more wait time before the engine is ready. That would also mean dropping gravity out of the picture.

I was just thinking a bot match would probably allow for more AI, and would make judging easy. It'd also be kinda funny to post movies of the matches...

I'm up for bot or gun, but I think I would like gun better just so Merri will have it done sooner and I can get working sooner. :)

Well may I suggest we decide how it'll work now (either way). Then those people that don't have a whole lot of spare time will have longer to work on it.

I'm pretty sure it will be the "gun" option just because that is what Merri has been generously working on so far.

Quote:

---

Originally Posted by Merri

You dislike programming?!?! A lot of it is math... dislike it and you shouln't be programming unless you are a masochist.

What did you think about the pathfinder contest? It seemed to be pretty popular and it is pretty close to this in the way that you are given just a simple interface, yet you need to figure out how to use that interface to the max to make it do something you want it to do. Believe me, figuring it out and making it work is a lot of fun and not "just" math.

So people seem not to be very interested about anything game programming related? Or just dislike me or my suggestion? Or both? Or all?

---

No, I hate algerbra. Anything math related is not my thing..I do fine programming though..

I have some doubts..

Quote:

---

* the gun would get information about the angle where the object is and how far away it is
* the gun would know where the object is heading to (ie. the end position)

---

Ok, having that, the speed can be calculated.

Quote:

---

* the object could vary its speed and path, but would be time limited to get to its target

---

This is interesting, Supose that we have a perfect AI that make a perfect shot, and while the bullet is in the air, the object change speed or turns, the shot doesn't hit the target, so it will be a LUCK factor here?
Also, All projects will be tested with the same Object path?

Quote:

---

* the gun would know the speed of the bullet

---

So the gun doesn't decides which will be the speed of the bullet, this is good, else the gun could make a super fast shot and that surely would hit the target.

Is the object running AI also, to detect who is firing the shot? :)

A super fast bullet would not necessarily make a hit, because it is so fast that it would "jump" on the field and that would make collision check to happen every X pixels, effectively skipping the X pixels in the middle -> if it makes no stop where target is, you get no collision detected and no hit. This is why I have made an "internal speed" feature in to the engine: not every change gets drawn on the screen so that the bullets can be slower, but have a much better accuracy. But it would still look normal to the user.

Quote:

---

This is interesting, Supose that we have a perfect AI that make a perfect shot, and while the bullet is in the air, the object change speed or turns, the shot doesn't hit the target, so it will be a LUCK factor here?

---

Then it is not a perfect AI! It is just a perfect target calculation. The main aim is to make the AI actually "think" about the situatation. If the target is flying at the same rate, then it is a good idea to shoot it before it flies away. Also, if the target is coming right at you, I'd shoot for sure as much as I can, because it is slow to change the direction. But I wouldn't shoot much if it stays far away and makes constant changes on where it goes, unless it is almost gone out of the area (you get information about this as well).

Also, there could be clouds or the target could go so far away you can't see it. You'd then have to use the information you have (you get information about how far you can see in each direction). Then it is just up to you how much intelligence you want to code: if a cloud comes, you could estimate how big the cloud is and then try to guess where the target could be.

And most interestingly, there is a possibility of making an AI that learns. You could make it feel good each time it hits a target. The engine can provide information on when the bullet that hit was shot, so that is then what AI learns to do better. If a bullet doesn't hit, the AI wouldn't do the failed thing as much. The rules wouldn't prevent submitting a premade intelligence file.

Quote:

---

Also, All projects will be tested with the same Object path?

---

Yes, I make a scripting feature which allows to make premade paths.



The stuff I originally listed won't be completely final information, maybe calculating an accuracy instead of pure amount of hits is better. Though accuracy isn't enough, because someone could shoot just one bullet. So a combined accuracy and number of shots could be used for judging and give points. Maybe points could be accuracy * number of hits.

1/1 * 1 = 1
2/3 * 2 = 1.333
7/10 * 7 = 4.9
9/10 * 9 = 8.1
30/60 * 30 = 15
43/60 * 43 = 30.8

Guess that is simple and fair enough.

Quote:

---

Originally Posted by dglienna

Is the object running AI also, to detect who is firing the shot? :)

---

I thought about making an AI for that for your amusement :) The engine is flexible enough for doing that.

I got an E-Card for my birthday asking me to click a button, for: SPORTSCAR, CASH, or GREETING.

Two of those moved around, making them impossible to click. I settled for the Greeting.

Call me a skeptic, but...

Sounds awesome Merri. This will be an awesome venture into game programming for me. How long do you think it will be until we have a version that we can start working with (IE - start the contest)? A week maybe?

Pino, is there anything we can do to help you run this contest?

For the contest of being running, I don't know. If Pino converts the code to VB.NET, that might take a good while of time due to my advanced coding techniques that might be hard to follow and that might cause a problem with VB.NET - I don't know for sure. Some people can get an early start: the engine might not be perfect, so it would be a good idea to push it out early so that people can give me suggestions on what more they need. I might forget to add an important feature and so on. I want to push it out early myself so the developing would ease on my part.

I probably need to write a simple documentation on what variables one can use...