Good Dungeon Design Principles

[SpiderFighter]

1. Amnesia
2. Cantilever

That covers pretty much every video game ever made!

[JohnWordsworth]

Some really good advice in this thread - it makes for a really interesting read. I don't want to re-iterate the above points by writing my own list here, but I did just want to mention what I think is possibly the most overlooked / hard to follow piece of advice...

Actually listen to your play testers
This is difficult, because it will include criticism and comments that require you to re-do your hard work. If you have 5 play testers and 1 comes back saying that a puzzle is too hard, you will be tempted to think "they just don't get it", but the data now actually says "it's likely that 1 in 5 people will get stuck at this point, then get bored and stop playing". Personally, I think what you really want from a play tester is honest advice (I liked floors 1,2,4 and 5 - but level 3 was boring), because then you can improve your mod. But you have to consider the feedback you get as "x in y people will think this". Play testing can do so much more than just give you a list of bugs.

I've looked at loads of indie dev prototypes in the past. When I comment on a prototype I try to give honest feedback packaged in a "s**t sandwich" - highlight a strong point, mention my biggest barrier to enjoying the game and then end on another high. Often I'm met with "ahh, you're just not getting it". Usually it's just something little - mechanic A needs to be faster / tweaked. I see 50% of others saying or reacting badly to the same thing, and the developer always comes back with the same response. Truth of the matter is - these people won't buy your game for the reason they told you, and the general public will likely think the same.

Organising a playtest over Skype is a great idea. An alternative would be to ask your play tester to record their game session (with their mic on if possible) as they play. This would reduce the chance of you saying anything to them. By giving them a hint to a puzzle or even introducing your mod in a way a general gamer won't get will skew your data dramatically.

[SpiderFighter]

Some really good advice in this thread - it makes for a really interesting read. I don't want to re-iterate the above points by writing my own list here, but I did just want to mention what I think is possibly the most overlooked / hard to follow piece of advice...

Actually listen to your play testers
This is difficult, because it will include criticism and comments that require you to re-do your hard work. If you have 5 play testers and 1 comes back saying that a puzzle is too hard, you will be tempted to think "they just don't get it", but the data now actually says "it's likely that 1 in 5 people will get stuck at this point, then get bored and stop playing". [...]

Excellent points, John. In retail, there's an old adage that goes something like this: Only 1 out of 10 dissatisfied customers will tell you they're dissatisfied and why. The other 9 will tell everyone else.

Play testing is essential. No matter how hard (or how many times you try) you will never think of everything. Play testers will purposefully try to break things and, while that can be frustrating to see someone trying to kill something you've created, it's necessary. There are at least a few mods (not necessarily for LoG) that I most likely will never play, simply because they were released too early, and were inundated with constant fixes after the release. Nobody wants to devote part of their life to playing something, then find out they have to do it all over again because of a game-breaker that wasn't discovered (Please note, I don't mean something that was missed during testing, I mean a mod that wasn't tested at all. Certainly not everything can be caught, but it does matter that we try.).

As for the Skype playtesting, Neikun organized a list that kind of fizzled out, but it'd be great to see it thriving, now that the community is starting to really see a lot of releases. List of Available Testers *Apply inside