Closed Account

[Decebalus_Rex]

Hi,
Closed Account

[Iluvalar]

8c-SG1max4v4_v1.wz there is a little hole in the cliff on south-west side or the circle.

You will have an oil distribution problem there: It is too easy to defend all the oil in the "middle". Once a team will get to defend that much the game will be already over. You will cause an urgent need to cover that area, and both team will rush there with their very first truck. This will cut all the more conservative openings where you start with a lab or with a couple more buildings out of the game. Without such strategies appearing once in a while, the game will quickly grow repetitive with the 2-3 possible strat repeated again and again.

Your layout look very interesting beside that. I dont want to discourage you, oil placement is an art .

oh ! And think about posting a screenshot next time

[Merowingg]

It is four tiles.. I would not call it small hole

I am also woried about about those very steep clifs which are made to be climbed.. (when units climbing over 45 degrees cliffs it will look unnatural/funny) there may also be while game confusion..

But in general the map is quite ok The way you use tiles reminds me of early me

[aubergine]

It's great to see a regular stream of new maps. I just wish I had time to try them all!

[Decebalus_Rex]

Hi,
Closed Account

[Decebalus_Rex]

Hi,
Closed Account

[Decebalus_Rex]

Hi,
Closed Account

[Iluvalar]

Hello Rex.

As i was trying to tell you in game, your maps have too much open spaces. This make moonwalking borgs and tanks that delay before they can move.

What happen is the path finding code, that try about every possible path on every possible tiles. since it's an exponential function. It goes quickly out of control.

Here is a picture of all the "dead zones" in your last revision. You could transform all this in thick block, and nobody would never see a difference in the map gameplay. This would cut half the map for path calculation.

[Decebalus_Rex]

Hi,
Closed Account

[Cyp]

Hello Rex.

As i was trying to tell you in game, your maps have too much open spaces. This make moonwalking borgs and tanks that delay before they can move.

What happen is the path finding code, that try about every possible path on every possible tiles. since it's an exponential function. It goes quickly out of control.

Here is a picture of all the "dead zones" in your last revision. You could transform all this in thick block, and nobody would never see a difference in the map gameplay. This would cut half the map for path calculation.

It's an exponential function? Ah, that explains the 1^1000 second delay on Sk-Rush and the 2^60000 second delay on NTW, when moving a tank.

The best for pathfinding performance is currently if the open area is convex, but I'm not sure pathfinding performance should be a large consideration when designing maps.

[Iluvalar]

It's an exponential function? Ah, that explains the 1^1000 second delay on Sk-Rush and the 2^60000 second delay on NTW, when moving a tank.

The best for pathfinding performance is currently if the open area is convex, but I'm not sure pathfinding performance should be a large consideration when designing maps.

Be handsome, get the true datas and stop annoying me.
http://en.wikipedia.org/wiki/A-star_alg ... Complexity

In big multiplayer games, the borgs start to "moonwalk". To be exact, there is a delay between the moment the walking animation station and the time were the borg start to move. The game become impossible to play correctly. I dont know how 3.1 manage this but in 2.3 the desynch just become awsome, and the whole map is impossible to play.

Thx Rex, i'll give it a try.

[Cyp]

...

Be handsome, get the true datas and stop annoying me.
http://en.wikipedia.org/wiki/A-star_alg ... Complexity

...

I'm not sure whether you're serious. We're not doing A* on the free group or anything like that.

The cost of pathfinding is at most approximately linear in the map area, or quadradic in the length of the path.

Or did you mean that the cost is exponential in the number of digits needed to write down the map area?

[Iluvalar]

The cost of pathfinding is at most approximately linear in the map area, or quadradic in the length of the path.

Quadradic then. But the risk is greater than that, because a bigger map have more obstacles. which increase the risk for the heuristic function to screw it up and the total cost go quadradic. So you start with a small map 128x128 with the cost moslty linear (1.2) for a total mean cost of 128^1.2=337 . To a big map 250x250 with mostly quadradic resolution (1.. with a mean cost of 250^1.8=20715.

Which create an effect by far worst than simply "quadradic". I dont know how you want to call that... but it's not linear at all.

[Decebalus_Rex]

Hi,
Closed Account

[Iluvalar]

I believe there is a "impassable" flag for the tiles. But do not bother with that. As long as you prevent to move in some zone or probably just limit the movement in some chokepoints, I believe you help the path finding a lot. My picture was just there to show you that some zone in your map was uninteresting and could be "removed" from the game area without lost.