Thargoids following player wormhole crashing into the buoy

[Commander McLane]

I've created a tweaked thargoidAI which allows some of the Thargoids to follow the player when he jumps out of interstellar space.

I've noticed a problem, though: In about half of my tests the following Thargoid emerged right on top of the witchpoint buoy, crashing both himself and the buoy.

Either my tests are suffering from a bad case of pseudo-random (if the witchspace exit position is determined by pseudo-random instead of random), or perhaps the safety-radius around the buoy, where no ship should be spawned, is too small for some of the bigger Thargoid models from Thargorn_Threat.oxp.

I also found myself spawned very close to the buoy sometimes (about 1000 meters).

I also observed (and already had a PM-exchange with Eric about it) that all following ships are spawned at the exact same position. I don't think that's necessary. It's especially annoying for Thargoids who don't move away from their original position.

Certainly the reasoning behind this is that one wormhole creates only one exit point, therefore ships following each other are all spawned in the same spot. However, while this may sound logically at first, it is actually inconsistent with current behaviour. First of all, the exit position of all ships is the same, but it is not the same as my own exit position. If they were following me, why don't they emerge at my position? Second, if I follow another ship in its wormhole, I don't emerge at its exit position, but somewhere else.

In other words: it appears that when following another ship, the first follower is not spawned in the original wormhole exit, but all subsequent followers are spawned in first follower's exit. If we want to convey a fixed exit position, this behaviour makes no sense at all.

Therefore I'd vote for dropping the idea of a fixed exit position completely. In ordinary gameplay, where there's usually only one follower, nothing hints to a fixed exit position anyway. Instead make it random for each emerging ship. Wormhole exits are fuzzy. That's why they don't leave a lasting blue disk, contrary to wormhole entrances, which do leave a lasting blue disk.

[Switeck]

Still happens:

when following another ship, the first follower is not spawned in the original wormhole exit, but all subsequent followers are spawned in first follower's exit.

I'm seeing lots of Thargoids follow me (due to my changes of thargoidAI.plist)...they tend to crash/bump into each other. If they're not destroyed, they tend to be badly damaged.

They also seem to lose half their bounty.
I've dragged a few back-and-forth between 2 systems till they were clean.

[Eric Walch]

I'm seeing lots of Thargoids follow me (due to my changes of thargoidAI.plist)...they tend to crash/bump into each other. If they're not destroyed, they tend to be badly damaged.

They also seem to lose half their bounty.
I've dragged a few back-and-forth between 2 systems till they were clean.

Most AI's are not prepared of really going through a wormhole, but if they do, they need special code to clear the addition area, like with launching ships.

And when ships would not loose half their bounty on jumping, it would a bug

[Fatleaf]

And when ships would not loose half their bounty on jumping, it would a bug

Just a side point. Shouldn't Thargoids keep their bounty. It seems daft that they could become clean.

[cim]

Most AI's are not prepared of really going through a wormhole, but if they do, they need special code to clear the addition area, like with launching ships.

escortAI.plist does seem to have that code, though

Code: Select all

"EXIT_WORMHOLE" =
         {
        ENTER = (
            setDestinationToCurrentLocation, 
            "setDesiredRangeTo: 7500.0", 
                        "setSpeedTo: 50", 
            performFlyToRangeFromDestination
        ); 

...and if I follow a ship + its 2 escorts through a wormhole, the escorts tend to collide on exit (not usually fatally {1} - but it can happen)

I think the problem here is that if you have 2 escorts following a mothership, and you follow all three through the wormhole - try to enter the wormhole *just* after the escorts do - then in normal flight the escorts will enter the wormhole at (near)identical times, so they'll also exit the wormhole at (near)identical times, which means that they don't have time to get clear of each other on the other side.

A km or 2 of randomisation in wormhole exit position for following NPCs would probably solve all the problems.

{1} I'm testing some modified escorts which complain when they take damage, so I notice this happening a lot, even though both survive.

[Kaks]

And when ships would not loose half their bounty on jumping, it would a bug

Just a side point. Shouldn't Thargoids keep their bounty. It seems daft that they could become clean.

Noted! Will be corrected post-MNSR: Thargoid bounty will be unaffected by jumps.

About Thargoids bumping into each other and/or the witchspace buoy, that issue should have been resolved in rev4675...

[Eric Walch]

One cause of the collisions is that all ships exit at 25% of their maximum speed. And because escorts have usually a higher max speed than their mother, they will fly in their back when following to close. After 1.76 we will probably change this in a way that all ships coming from one wormhole will have the same exit speed. I think that also will largely solve the problem.

However, I must say I never witnessed escort crashes and I also did some testing for 1.75. The current escortAI already has a new "EXIT_WORMHOLE" state that largely improves things in comparison with pre 1.75 Oolite. Problem I had is that escorts have lost their mothers escort group on jumping and negotiating with the mother takes time leading to crashes. As it is now, they should fly in a random direction for 7500 meter and than start negotiation with their mother.

One other problem is that the current AI has to deal with very different situations. e.g. When the mother jumps and the escorts follow, the mother will always be in place when the escorts arrive. But when the player jumps out of the system and a pirate with escorts is following the player, it is likely that the faster escorts jump sooner than their mother. So, the AI needs a waiting time after a jump before escorts can decide that they completely lost their mother. Both situations were tested before I added my changes. I could write better code for one situation, but than the code would fail for the other. This was the best compromise I could make without a full rewrite.

[Switeck]

It is this action in thargoidAI.plist that causes them to sit still after emerging from a wormhole:

Code: Select all

"LOOK_FOR_TARGETS" = {
ENTER = ("setSpeedTo: 0.0", performTumble, scanForNonThargoid, "pauseAI: 5.0");

The performTumble does an instant stop from whatever speed it initially has. And if that wasn't enough, the "setSpeedTo: 0.0" and "pauseAI: 5.0" makes sure it stays that way for a bit.

(Somewhat unrelated) I think the line

Code: Select all

"NOTHING_FOUND" = ("setStateTo: HEAD_FOR_PLANET");

should probably be

Code: Select all

"NOTHING_FOUND" = ("setStateTo: HEAD_FOR_WITCHPOINT");

...to get around the problem of interstellar space not typically having a planet. This should also make Thargoid groups less likely to diverge away from each other.

[Kaks]

Well, we do need something a bit more elaborate: they still need to head for the planet in order to find victims when not in interstellar space...

[Eric Walch]

(Somewhat unrelated) I think the line

Code: Select all

"NOTHING_FOUND" = ("setStateTo: HEAD_FOR_PLANET");

should probably be

Code: Select all

"NOTHING_FOUND" = ("setStateTo: HEAD_FOR_WITCHPOINT");

...to get around the problem of interstellar space not typically having a planet. This should also make Thargoid groups less likely to diverge away from each other.

?? As far as I know does all this already happen since Oolite 1.74. And when not, it definitely was in 1.75. When in interstellar space, thargoids will not try to go to a planet but to the witchpoint.

And thargoids hitting each other when leaving a wormhole is no bug as they have no AI to go through wormholes. And when you change the AI to let them do that, you must change the whole AI to deal with this situation. It is no bug in the official AI.

[Switeck]

The Thargoids diverge apart if heading to a planet. Any Thargons they launched get left behind and go inactive. Even if there was a large bunch of Thargoids, before long there's seldom more than 1 Thargoid within a 25 km radius. Some even suicide on the invulnerable main station. It's a joke to kill them at that point.

I've had Thargoids run into wormholes left by other npcs even if they didn't "intend" to. Admittedly, it was a lot of work to trick them into entering a wormhole. I don't need OXPs to do that, it's just much easier that way.