Follow by Email

Friday, 30 January 2015

Development Blog #0010 - Server Population

If you don't know Wool, click here.

This week I made newmans log on and off the server.

Newmans will now go back to their base and log off after they have played for between 2.5 and 3.5 hours.  If you watch this week's video, you'll see newmans disappear after they have logged off.  Newmans will come back online at the same time each day and play.  To keep the model accurate, I updated the AI and the combat system so that newmans will attack sleeping enemies without hesitation, and are guaranteed to win combat.  Cool, now I've got a baseline model of Rust!

Wool Devblog 10 - 5 Days @ 20x speed

Server Population
I got in touch with Bugs, the guy that runs Rustafied. Using some estimates he gave me about the server base, I've programmed Wool to model the average server.  The Wool server sees a peak of about 60 players at 12:00, and a low of about 20 at 0:00.   I also looked into Steams stats for Rust, and to no surprise, found a bell curve player distribution.  So I programmed this trend into Wool's server.

Online Steam Users

Raiding is not common, but is possible.
If you watch for Garry's base in this week's video, you'll see that it survives about a day and a half before being raided.  A couple people managed to build a tier 2 base before their ~3 hours were up, and all of those bases survived the first 24 hours.  However, some people with tier 1 bases did get raided, and they'll wake up dead and poor.  I'm going to change things so that over 50% of people get a tier 2 base for three hours of work, because that's what I set out to do.  I'll probably tweak the base cost to get the desired result.

There are bugs affecting the AI.  I've programmed AI logs to see where they are making the wrong decisions, and it's going to take time to sort though that.  However, most of the time, they're quite capable.  I'll make this a priority next week.

Newmans are not aware of other newmans unless they run into them.  I'll update the AI to make intelligent fight / flee decisions.  I'm thinking that, since most people KOS, I'll make newmans KOS if they have equal or better weapons and armor than their target, and flee otherwise.  Newmans will flee for a random duration before deciding to turn and fight, and conversely, their pursuer will only chase for so long too.  This should be done next week.

Data Mining
After the AI is in good shape, I'll draft up which graphs I want to show, program the data collection for it, bring the collected data into Matlab, manipulate it, and finally display it.  These graphs will provide evidence that this progression system is fun to play, and well balanced.

No comments:

Post a Comment