About sonelliot


Ludum Dare 34
Ludum Dare 33

sonelliot's Trophies

sonelliot's Archive

Guard-Master Post-Mortem

Posted by
Friday, August 28th, 2015 8:48 pm

For our first LD we created an inverse-stealth game called Guard-Master. You can play it here. The core idea was to create a typical stealth game where you control the guards rather than the intruder. Simple right? Well not quite.

What went right

We started the jam with an epic 3-hour brain-storming session. Peter had read about the 100-10-1 rule on Reddit and we were keen try it out. You start by coming up with 100 simple ideas and then culling them down to 10 of the best (and doable) ideas. Finally we selected one of the top 3 ideas that we felt was easiest to get done. We liked one of the other ideas so much that we’re actually prototyping that one now as well. You can checkout our list of ideas here. Be warned though, some of them are very outrageous! I’m really glad we took the time to work through this process. In past jams I have been in teams where we jumped on an idea way to quickly, just because it was easy to visualize and then regretted that decision later.

The other thing that worked well was creating the levels with the Tiled map editor. At first we were worried that we wouldn’t be able to find a level editing tool for the perspective we had chosen. As it turns out Tiled not only supported it but also exported in a format that Phaser could load and display easily.

Finally, something that I think we did well was the aesthetics. We chose a simple pixel-art style for the characters and tileset. Peter did a great job creating all of the environmental bits and pieces. Visually everything came together quickly. In fact we had a lot of props that didn’t make it into the final game because we didn’t have enough time to code-up their behaviour.

What didn’t go right

We’ve read over and over again that you should prototype the minimum viable product of your game first. Everything should be coloured boxes and all you should care about is the core gameplay. I’d be lying (a lot) if I said we did this. We didn’t really have anything playable until the end of day 2 and that would be being generous with the term playable. It was only once we started playing our game that we realized how not-fun it was.

In our original concept for the game, the intruder would try and sneak his way to the exit without being detected. We also wanted him to incapacitate unsuspecting guards along the way. I spent ages tinkering the A.I. to try and get this to work, but it was a bit too hard. More often than not the intruder would just wander into the guards field of view and be easily disposed of.

Eventually we scrapped this idea and went with a much simpler hide-and-seek approach. The intruder will now just pick the furthest hiding spot from all of the guards and if you don’t find him within a minute you lose. This approach was a bit more fun. We also changed the guards from auto-firing to manual shift-click fire. This made it a bit more of a challenge to hit him once he had been discovered. It is still a bit broken though. Since the intruder always picks the furthest hiding spot, he will predictably be in the same place once you reload the map. Bah, well at least the first playthrough might be fun-ish :p.

Final thoughts

Overall we had a lot of fun. It is also really encouraging to see the feedback we’ve gotten. Even though we made a broken game, people seemed to like the idea and could see its potential, which is all we could really hope for. The biggest lesson that we’ve learned would be to always prototype the minimum gameplay first! 

[cache: storing page]