Probably an issue caused by the walls. When using moveTo creeps will repath every time they enter a room. When pathing from within the room you're trying to claim it can "see" the walls and will attempt to avoid them. This path may then take it back into the original room where it'll repath while not being able to "see" the walls. This can kick you into an infinite loop.
Lots of possible solutions. Using maxRooms: 1 as an option when you're in the target room is probably the easiest.
What I think went well
findRoute don't have a clean way to hook into them as the block in the middle of a room.What I think did not go so well:
I'm looking forward to season 3. Balancing time between that and arena may be tricky.
@o4kapuk said in Season #2 is open!:
World sectors are isolated from each other by huge constructed walls on the season start, so you have to dig your way to distant players
Any details? How huge? They don't appear to exist yet.
Collecting score resources and fighting over score resources is a good CPU/energy sink. Not sure we need anything else.
Hopefully both, depending on the arena. Seems possible to make it optional without too much design headache.
@gadjung said in Season #2 concept:
@tigga said in Season #2 concept:
I still think the best way to achieve competition across all levels is to make the world non-uniform so there are good regions where strong players will want to be, and worse areas with less reward where lower level players can only have other low level players as competition.
For non-uniform world, my issue is that it leaves bigger parts of it less-usable and discourages respawning into 'worse' zones. Weaker players are not much affected by that since they don't have codebase with features that would get impacted by that. For 'medicore' players that starts to be a bigger impact.
So from good player perspective 'Center' makes sense, for weak-player there's no difference, for medicore one it's either center/close to center and then respawn into worse area after wipe, or just start in worse area and have a boring/limited play
Why would it be boring/limited? You'd still be competing against others in your neighbourhood and it would be more fair competition and more likely to be fruitful as you're likely competing for similar scoreboard positions.
@bogden said in Season #2 concept:
As far as new player goals are concerned, I think seasonal is a wonderful format to get people started in a friendlier and safer environment without totally messing up the core gameplay of regular screeps.
I think this is an interesting point. I think the opposite: seasonal should be a competetive environment compared to the tame MMO experience that most people have. MMO is friendly and safe. Seasonal should be more competetive.
The challenge is to have it competative at the high end and at the low end. I think season 1 struggled there. It's not yet clear to me whether the rules proposed will change this: eg. a strong player has little incentive to keep a weak player with a score type the strong player (or a strong players' alliance) already owns alive. The rules heavily favour pre-season organisation, which I think leaves a lot of weak or new players behind.
I still think the best way to achieve competition across all levels is to make the world non-uniform so there are good regions where strong players will want to be, and worse areas with less reward where lower level players can only have other low level players as competition.
@bogden said in Season #2 concept:
The reason I say this iteration on seasonal may be a step in the wrong direction is due to the added element of required social interactions with other players, meaning diplomacy, deal making, backstabbing, and more. Screeps is primarily a game for programmers, by programmers, and I believe that these forced social elements will further deter new players from joining the game.
Or old players from joining the season. Personally I feel a player's performance should be measured by how well their code performs, not how many freinds they have.
IMO most of the calls for co-op in the previous thread were calls for global co-op, where everyone cooperates toward a single goal and there are no winners or losers. I think that might be interesting. Team based eco-wars, which sounds like this is heading toward, sounds a lot less interesting to the point whether if the rules outlined in the OP were used I'd be at most 50-50 on participating, probably leaning toward not bothering. If a pre-made team of 5 players beats a team of 3 players by virtue of having 5 players instead of 3 that's not really very exciting. The game scales too well with player-count-on-team (GCL, CPU). It would have on season 1 as well had it gone that way and I think it's a step in the wrong direction to encourage rather than discourage this. For example: if all YP players decided to co-operate I think we'd see YP players 1st and 2nd. That'd be decided before the round even starts.
So, global coop would be ok. Some other system to allow new players to participate safely. eg. as suggested in the thread, a world with non-uniform scoring resource distribution so the top players would be forced together leaving the larger and less valuable areas to newer players. Forcing co-operative co-operation is not good IMO.
@rayderblitz said in Ideas for Season #2:
@donatzor said in Ideas for Season #2:
Paint the World
- Each player is assigned a color value.
- Either signing a controller or 'moving onto' a room position 'paints' that room/roomPosition the player's color
- mapVis shows painted rooms/positions
- Player with the most 'pixels' (roomPositions) or rooms painted wins, can score from 'most covered area'.
- Could also 'paint' with energy and have some low eng to build walkable structure as a 'claim' on that room position as well.
I honestly love this idea. It'd be so visually pleasing to have whole rooms just painted in your color alone, but it'd also be a nice take on the gameplay in general, i think. I didnt know until now, but i really want this to be a season theme.
You'd have to be careful not to make the season goal irrelavent until the last few days. Tying it to the controller might acheive that (claim, sign, wall off controller, defend). Or accumulation of score over time based on signed rooms.
Firstly, I'd like to say I'm a huge fan of this season so far. And the concept in general. Screeps from GCL 1 is much more fun than MMO based screeps. It's also a bit my full-on though.
With that said maybe it'd be nice to have a break rather than back-to-back seasons. At least a two week gap between them, or maybe a month? I think you risk burn-out running them too frequently as I feel players put a bit more effort into seasonal than regular Screeps and that's hard to sustain non-stop.
I have some feedback on the current season which I'll share and I guess might help inform the next season.
Collectors and containers seem to be fairly well balanced. To compete for top spots you need multiple collectors. Collector capacity is low enough to prevent "bursting" meaning fighting for collectors is effectibe. Regen is about right for the score generated in the region. If you expanded and took land and more collectors it is easier to score high. This is how it should be.
Game start wasn't handled perfectly: the rules dropped very late and so the collector:container balance was unknown to players until after they'd spawned. This meant it was impossible to make good decisions, and I think that's bad. For example: should I spawn near the centre crossroads with four collectors, or should I go for the best room I can find across the map.
Some people struggled to respawn after dying - things got fairly cutthroat. One suggestion on slack was to explicitly stratify the world. Make score spawn more (and collectors regen faster) closer to the centre of the map. Maybe bias the map generator so rooms closer to the centre of the map are more likely to be two source rooms, or less likely to be swamps. This creates stratification: good players want to be closer to the middle. Weaker players can then spawn at the much safer outskirts. There's a risk:reward tradeoff: the winner won't likely be on the outskirts, but at least you can play.
No market worked fine. We're not even done with the first month and there have been multiple T3 boosted fights. I think no terminals would be fun, but more game changing. Not really sure.
@bodgen suggested it might be nice to be able to attack a collector. Maybe using claim parts you could attack it and gain a (very) small fraction of score placed into that collector, deducting from other players. Not a very "Screeps" thing as stuff always goes down. Does allow for more combat options around collectors.
Total tick count and 100 CPU gives a fairly nice balance between sub RCL 8 play and RCL 8 CPU conrained play. I do think min(100, 20 + 10 * GCL) would work fine as well.
Docs say "Attacks all hostile creeps/structures"
Walls are not hostile - they are neutral.
@kingkong said in What is up with shard0? The unfairness of slow ticks.:
Is it a fair conclusion that competitive play and cleaning out zombies is part of the game design as way to get better tick times?
This is just a question to better understand the game, not a comment in the form of a question.
I'm not really sure it's part of the game design. It just is what it is.
Try passing range: 1 as an option - you don't actually want to move to the source - you want to move within 1 tile of it. Your current code will try to find a path into a wall, which causes issues.
@rhef said in A release procedure for changes:
which was running at 5k for years without problems
I assume that's hyperbole? Pixels were launched (exactly) 6 months ago!
I agree with the general sentiment though IMO most changes have been fairly well telegraphed recently.
@tdxtor said in Is Screeps ever receiving a Discord?:
@shibdib I feel like a lot of people have trouble even knowing about the forum's existence and even fewer know about the slack. I'm surprised there's no link to it in-game.
In game there is a link to the forum and to slack.
@o4kapuk said in Season #1 announcement:
We decided to leave game API as it is because removing the source of credits looks enough for disabling the entire market.
Presumably with send auto-blacklisting other players?
Hopefully the official grunt can be updated for seasonal https://github.com/screeps/grunt-screeps (and hopefully this PR merged at the same time: https://github.com/screeps/grunt-screeps/pull/19 )
It's a simulator, not the real thing. This reload is one of the things where it differs from "real" servers.
https://docs.screeps.com/api/#Game.market.incomingTransactions
[{
transactionId : "56dec546a180ce641dd65960",
time : 10390687,
sender : {username: "Sender"},
recipient : {username: "Me"},
resourceType : "U",
amount : 100,
from : "W0N0",
to : "W10N10",
description : "trade contract #1",
order: { // optional
id : "55c34a6b5be41a0a6e80c68b",
type : "sell",
price : 2.95
}
}]
What's wrong with amount * order.price?