Respawns in areas that you aren't in

[Beirlis]

I recall reading about how respawns work somewhere, each monster has some random chance to respawn for each second (or turn, not sure) that you occupy that area, but if you are away from an area for more than x seconds (I think it was 300-ish) or you rest, then all mobs in that area will respawn. I would have to say that this is flawed. I think that the game should manage respawns of all areas that are not fully populated so that the random respawns occur even when you are not in that area. The result of not doing so is that you clear area 1 and move to area 2. You clear area 2 and move to area 3. You clear area 3 and move back to 2 and 1, but both are fully unpopulated -- they should have had a chance to respawn *some* of the monsters there. But now, since you have re-entered each area, the all-mobs-respawn timer has to start all over again.

Alternatively, if you don't want to bother with the overhead of managing monsters from multiple areas at once, then when you re-enter an area that isn't fully populated, run a respawn, but multiply the chance for each mob to respawn by the quotiont of the number of seconds you were gone and whatever that all-mobs-respawn timer value is. So if you're gone for 180 seconds, and the all-mobs-respawn timer value is 300, then give each missing mob a 180/300ths (60%) chance to respawn as the player re-enters the area.

[Zukero]

That's a very valid point !
In your example, I'd rather try the respawn up to 180 times (for the 180 1-second periods that elapsed) with 1% chance (like the 1-second tick when you're on the map) for each monster. It would be statistically equivalent to the behavior if you stayed on the map.
Maybe rijackson could come up with a more elegant formulae to have a statiscally equivalent test for x tests with 1% chance of success in a single computation.

I'll bring this up in the next dev meeting.

[fiernaq]

It's 180 seconds for a complete respawn (3 minutes) but yeah, I agree that this system is faulty/not-very-good.

[PK17]

It might make the first 5-10 levels a little more challenging but after that it would be more than welcome amongst most player. +1

..... I don't see why this is a bug, how about a move to ideas and future development.

[Beirlis]

It might make the first 5-10 levels a little more challenging but after that it would be more than welcome amongst most player. +1

At least the penalty for dying isn't too high then yet. Plus, if you know how it's going to respawn, you can be more careful about how deeply you get yourself into a given mess.

..... I don't see why this is a bug, how about a move to ideas and future development.

I would call this a flaw or bug in the design as the curent behavior is counter intuitive and the proposed is much more intuitive.

[rijackson741]

I recall reading about how respawns work somewhere, each monster has some random chance to respawn for each second (or turn, not sure) that you occupy that area, but if you are away from an area for more than x seconds (I think it was 300-ish) or you rest, then all mobs in that area will respawn.

If you are in a place there is a 1% chance per respawn area that a monster respawns every 0.5 seconds. All monsters are respwaned if you don't visit an area for 3 minutes. See here: http://andorstrail.com/viewtopic.php?p=588

I'd rather try the respawn up to 180 times (for the 180 1-second periods that elapsed) with 1% chance (like the 1-second tick when you're on the map) for each monster. It would be statistically equivalent to the behavior if you stayed on the map.
Maybe rijackson could come up with a more elegant formulae to have a statiscally equivalent test for x tests with 1% chance of success in a single computation.

The probability of a respawn in one respawn area is P=1-(1-p)^n, where n is the number of attempts, and p is the probability per attempt (in this case 0.01). So calculate n as time * 2, then calculate P (time is in seconds of course). Then generate a random number between 0 and 1. If the number is less than P a monster respawned in that respawn area. After 3 minutes the probability that a monster respawned in a given area is 97.3%, so I think it would be a good idea to then set it to 1. That way you don't have the bookkeeping overhead of always having to keep track of the elapsed time since visiting every place in the game.

It might make the first 5-10 levels a little more challenging

It would certainly do that, yes. When a player first enters a place, the situation would be the same as it is now. After leaving a place for 3 minutes it would also be the same as it is now. But within that 3 minutes there would be a lot more monsters. Here's what the probability looks like as a function of time:

Respawn prob.png

Note that after only one minute there is a 70% chance the monster respawned, and after 2 minutes the chance is greater than 90%.

A possible way to re-balance the difficulty would be to drop the requirement that the statistics are the same whether or not you are in a place. Make P a simple linear ramp, rather than an exponential function, which starts at 0 at time 0 and ends at 1 at time 360 seconds (6 minutes). Then at 3 minutes, on average, half the monsters will respawn when you reenter the place. At 6 minutes they will all respawn. That will not change the average number of monsters compared to the way it works now, it just evens them out over time, so it will eliminate the problem Beirlis brought up.

Edit: I just realized that what Beirlis suggested is also a linear ramp. The only difference is that I am suggesting the ramp go from 0 to 6 minutes, rather than 0 to 3, so that the average number of monsters is unchanged from the way it works now.

[PK17]

It might make the first 5-10 levels a little more challenging but after that it would be more than welcome amongst most player. +1

At least the penalty for dying isn't too high then yet. Plus, if you know how it's going to respawn, you can be more careful about how deeply you get yourself into a given mess.

..... I don't see why this is a bug, how about a move to ideas and future development.

I would call this a flaw or bug in the design as the curent behavior is counter intuitive and the proposed is much more intuitive.

If that's your logic okay, it was designed to be that way, even if it is in need of a revamp. Is every concept you don't approve of a bug?

[Zukero]

Thanks rijackson ! Just what I expected from you except it's even better !
Two questions : why is n time * 2 and not just time itself ? which p to use if we want P (n=180)=0.5 ?

I like the idea of keeping the consistent exponential ramp. The beginning area needs a rebalance anyway... (more quests for example).

[rijackson741]

Thanks rijackson !

No problem. Glad to make some contribution

Two questions : why is n time * 2 and not just time itself ? which p to use if we want P (n=180)=0.5 ?

1) Because, according to the post by Samuel in the thread I pointed to, when you are in a place there is an attempt to respawn every 0.5 seconds, so there are two attempts in 1 second.

2) Rearranging the equation, p=1-(1-P)^(1/n). So if we want P to equal 0.5 at 3 minutes (when n=360, not 180), then p=0.0019. If we round that up to 0.002, then at 3 minutes P=0.51. At 15 minutes (n=1800), P=0.973. Here's the graph:

Respawn prob2.png

This might be a good compromise

[fiernaq]

I'd think a better in-game result would be to just flip that curve over. Gradually increasing rate of respawn that rapidly accelerates as time progresses rather than quick rate of respawn that slows down as time progresses. That way it would behave much more like it does right now except that you wouldn't have to wait in a different room for 3 minutes because even staying in the same room for 3 minutes would still almost guarantee every monster would have respawned.