2019
01.17

Since releasing the Paradice Five Hundred game all those years ago, one of the most consistent pieces of feedback I receive is that the AI partners and opponents don’t play optimally.

This.. is very likely true. I’m limited in that I can only design a computer player that plays as well as I know how to play, I’m not the best player by any means, and to be honest, I don’t have any group of players in my current city that I could even play with to get any better.

Previously, commenters have offered to assist me in improving the AI, but the advice has turned out to be in the form of vague statements involving things like ‘common sense’ or “why doesn’t it do X” which are, without the context of the hand, are quite simply not specific enough to program! So I would like to issue a challenge. I’ve created a Decision Tree diagram that represents how the AI currently makes it’s decisions; here it is:

Paradice Five Hundred – Card Playing Logic (click to enlarge)

Help me make this better!

If you’d like to play Five Hundred against computer opponents, but you’re frustrated by the way they currently play, I encourage you to click into and scroll around the current diagram.

Where you find errors in how the AI plays – feel free to write create a new section of logic (ideally draw it as a similar diagram, but I’ll settle for anything readable) and make it really clear EXACTLY where in the decision tree you’re replacing, and make sure every scenario from that point downwards is covered. Once I receive something, I’ll update the back-end code and release a new version, which you can then play!

To make the above decision tree, I used a tool called SmartDraw (free 7 day trial online), but really, you could use Powerpoint, Excel, or even just use Paint and edit the existing image.

If you’ve created something, leave a comment below with somewhere you’ve uploaded the image, or just your email address and I’ll get in touch.

1 comment so far

Add Your Comment
  1. Lovely content keeps it up helpful for the public, good continuation; I’m subscribing to your feed.