PC Alpha 19 Dev Diary

Status
Not open for further replies.
I recently decided that traders currently bring more bad than good to the game, so I started playing without them. When we have more fleshed out NPC, White River clan, bandits, better quests, then traders would be justified. Currently they bring weird problems/mechanics:

- the concept of invulnerable blocks/POIs

- the magic POI restoration/loot respawn and gimmicky double POI looting, which became a new standard

- the magic teleporting after 22:00

- the 'problem' of too many foods in store, 'solved' by introducing magic 'candies'

- exploits of 'secret stash' (putting on/removing nerdy glasses)

- the current quests we have are beyond primitive, even by sandbox standards

- traders/quests make the game easier, so people often see no point to build a base anywhere but next to a trader

- the opportunity to buy the things you need, instead of looting it, results in a scheduled routine of visiting every trader after the restock cooldown.

We can argue about the last point, as having more than one way to gain things enriches the game. Maybe yes, but I decided that I prefer the old way of looting/crafting. So when I found that Nitrogen has an option to build maps without traders, I did it. But you also need to use a modlet to remove the 'find a trader' quest with this opiton.
Not disagreeing with you as you make some good points.  Keep in mind RPG elements and a story has been planned since the beginning and the trader and the quest system are all important cogs in that long term goal.

It's difficult to have a semblance of progression for a story arch without a structured system to guide the player through it.

If anything it goes the opposite direction to open world non linear gameplay.

I commend the devs for trying to integrate both together into something that will be fun for as many as possible at the end of the day. 😀

Edit: Regarding buying loot, were you around when there were no traders? ForgeAheadBook gate?  Traders are not perfect but provide an alternate path instead of relying on loot RWG which was too one dimensional.

 
Last edited by a moderator:
B* is a search algorithm, it's just not very common, and not really suited for 3D path finding.

It's a really cool algorithm and I'd highly recommend reading it. https://apps.dtic.mil/dtic/tr/fulltext/u2/a059391.pdf

It's best used for optimizing decision making, and it works equally well for an optimality search as well as an adversarial search. It 'could' be used for 2D/3D path finding, however the special sauce that B* brings is a pair of values for each node. One is an optimistic search cost and the other a pessimistic cost. These values represent the bounds for how expensive a path that traversing through this node is. As the graph is explored, these values are updated when more accurate data is found. This can be beneficial if a nodes optimistic search cost is higher than the pessimistic cost of another, it can be excluded entirely because there is clearly a better path through the second node. This is really cool, but I cannot think of a good way to weight these values initially. Since the player can arbitrarily build structures, the pessimistic value would have to default to infinity, or some arbitrarily large value like 4 blocks of steel blocks. The optimistic value would be the difficulty of traversing the block contained in the node. So none of the special sauce B* has does anything except add overhead.

Further, I can't figure out a way to integrate the X,Y,Z heuristics like we can with A*, or similar greedy algorithms.

I'd love to see how B* could be adapted to 3D path finding, but I really don't think it's possible.
Huh, that's a new one on me, well done.  It does seem as though the performance would be bad even next to A*'s clunky pace.  

Why not move into the millenium and use a navmesh?  You'll be using an A* variant behind the scenes in places but in the era of modern computing there's little to be gained from using outdated and slow algorithms.

 
Last edited by a moderator:
Huh, that's a new one on me, well done.  It does seem as though the performance would be bad even next to A*'s clunky pace.  

Why not move into the millenium and use a navmesh?  You'll be using an A* variant behind the scenes in places but in the era of modern computing there's little to be gained from using outdated and slow algorithms.
A navmesh in a dynamically changing voxel world, that has nicely defined block-shapes?

That would not be a senseful choice.

(prebaked) Navmeshes are better used in static levels.

 
A navmesh in a dynamically changing voxel world, that has nicely defined block-shapes?

That would not be a senseful choice.

(prebaked) Navmeshes are better used in static levels.
I'm not on about using the built-in unity version, they're not that hard to create from scratch.  What would be so bad about recalculating a section of the navmesh that wouldn't be equally bad for an A* implmentation?  You have well defined block shapes which makes polygonising the terrain pretty damn trivial.

The use of half-blocks negates any advantage of a grid.

 
Last edited by a moderator:
I'm not on about using the built-in unity version, they're not that hard to create from scratch.  What would be so bad about recalculating a section of the navmesh that wouldn't be equally bad for an A* implmentation?  You have well defined block shapes which makes polygonising the terrain pretty damn trivial.

The use of half-blocks negates any advantage of a grid.
You not only need the freely passable areas, but also a search though any obstacle, given its resistance (blockdamage costs). That needs to be know for all potential voxels in the area. I dont see a navmesh having any advantage here, compared to using the raw 3D voxel data, makinging some custom edits (to detect jumpable gaps for example), and running a pathsearch on that.

 
You not only need the freely passable areas, but also a search though any obstacle, given its resistance (blockdamage costs). That needs to be know for all potential voxels in the area. I dont see a navmesh having any advantage here, compared to using the raw 3D voxel data, makinging some custom edits (to detect jumpable gaps for example), and running a pathsearch on that.
It's not that either has an advantage on the calculation speed of the route.  With the grid-based A* we currently have zombies following chessboard movement where it's advantagous to align yourself with the world grid and quite a few exploits explicity derived from half-blocks appearing to be full-blocks to the AI.  Neither of these would be an issue with a navmesh and accounting for these loopholes is going to end up costing you more system resources with A*.

 
Because its grounded in reality? Do you not hear of "tainted supplements" in UFC? Supplement makers pump drugs into stuff all the time to get people to actually take it. Plus its in the future. Look at 5 hour energy. Its quite possible/plausible for candy/supplements to get added benefits in the future as a marketing ploy just like 5 hour energy. Why do you get more resources when you mine with rock crushers? Placebo? Maybe, but they jam a cocktail of stimulants in there and you go hit that @%$*#! harder than normal because the suggested effect and stimulant made you want to get a better result than normal.
Not a bad argument tbh, maybe ill end up giving them a shot.

 
also, speaking of vehicles, when i close a door, the motorcycle / minibike / bike fall over :') is that something that is being worked on as well? i know it is a super minor thing, but it caught my attention the other day.
Working as intended. They don't have kickstands.

 
I actually do use the various buff items.  Every horde night, if I have recogs and learnin' elixirs, they're on my hotbar.  When I was doing a brawling playthrough, I happily brewed and guzzled beer and moonshine.  I love me some Megacrush, and if I'm encumbered, I will use steroids.   Recog is especially amazing because as long as it lasts, you're getting several bullets worth of killing power out of every bullet, so you use a lot less bullets.  I look forward to being able to just outright buy buffs I need out of vending machines instead of hoping against hope to find a rare recipe or spending several perk points to learn it.  

 
In A18 when you use a 2 block ramp as a steep wall 4 blocks high you are untouchable. Zombies don´t hit the wall, jump against it and slide down slowly... forever...  did you fix that in A19?
Video?

. . .Wrench cars my dude.

Gas is so ridiculously trivial I didn't know people even ran out. Less than 5 minutes of mining would get you a lifetime supply of it for 5 people, but if you absolutely refuse to even touch mining, you should already be wrenching cars you see fairly often and each one will get you hundreds of gas.

Even my "run and gun only" friend always carries a wrench with him on road trips so he can stop and wrench down a car and top off his bike if he gets low. Literally, 45 seconds of wrenching will run your bike for 10+ minutes
The 4x4 sucks quite a bit of gas, you can't harvest enough gas IMO, you have to mine shale. Maybe with scavenger perk maxed you could get enough fuel. I built one this game but I can't craft gas yet so I'm still driving my motorcycle.

 
Do you guys ever monitor the Subreddit? As with basically all games, the subreddit is like 10 times bigger than the actual forums or the steam forums and seems to be the main place people discuss the game. Like me for instance, I've played for like 3 or 4 years but never even really thought about posting on the official forums until the A19 signup stuff was here. The subreddit is quite active, especially after big updates
I don't. I'll bookmark that. Thx

 
Video?

The 4x4 sucks quite a bit of gas, you can't harvest enough gas IMO, you have to mine shale. Maybe with scavenger perk maxed you could get enough fuel. I built one this game but I can't craft gas yet so I'm still driving my motorcycle.
I thought you were going to speed it up to match the motorcycle and nerf gas guzzling a little?

If people don't use it for some reason or other, it's a wasted asset.

We used two once in mp, and everyone took thier motorcycle back because the storage of two motorcycles was about the same, and it was just so much faster to get back.

(It was nice though to use the passenger with thier map open to navigate at night.)

 
I thought you were going to speed it up to match the motorcycle
They did. In our MP save my wife follows me in the 4x4 while I'm on the motorcycle. As long as I don't keep threading through narrow gaps between objects that stop vehicles it keeps up perfectly, even with holding sprint and nuking the fuel economy on both vehicles.

 
Status
Not open for further replies.
Back
Top