Three in a Rogue - Match3 meets Roguelike

[NotExplosive]

I made this game for GMTK jam and have since added one major content patch that fixed up a bunch of stuff, added enemies, and a map editor. I plan on adding more stuff too but I figured I'd share it here to get feedback.

Made in LOVE 11.3, any 11.x should work though.

https://notexplosive.itch.io/three-in-a-rogue-classic

Would love to hear your thoughts!

[randomnovice]

Looks good! It took me a little while to understand how the match-3 mechanics worked and that they only happen when there are enemies around to defeat, but then the game became a lot of fun! Well done!

A few extra thoughts for improvement...
- As far as I can tell, there's not a lot of difference between X and Z attack when the level is empty?
- I didn't play long enough to discover if there are levels or if its truly just endless? It might be nice to have some sort of level structure and a minimap perhaps?
- Or another way to incentivise keeping playing would be a way to power-up your abilities or gain extra abilities?
- You might be happy with the randomness of the match-3 mechanic but if we knew at least the next row that was going to fall it would allow planning a bit more for strategy. Otherwise when you match a column you just have no idea what moves will follow!

[Davidobot]

Super cool! I really love the mash-up!

I have an unhealthy habit of testing love2d games and whether or not they work with the love.js 11.3 web port I helped with, and this one does! Since it goes into fullscreen, you probably should maximise (F11) your chrome window and then click the "go fullscreen" button at the bottom of the page, but it works well!

I thought this might be useful for you to know as jam games that have web versions usually get a bit more plays.

.zip of your game in web format:

three_in_a_rogue_web.zip

(6.42 MiB) Downloaded 249 times

(will only work on Chrome when uploaded to any webserver, or served locally)
Link to web port: GitHub

[NotExplosive]

Super cool! I really love the mash-up!

I have an unhealthy habit of testing love2d games and whether or not they work with the love.js 11.3 web port I helped with, and this one does! Since it goes into fullscreen, you probably should maximise (F11) your chrome window and then click the "go fullscreen" button at the bottom of the page, but it works well!

I thought this might be useful for you to know as jam games that have web versions usually get a bit more plays.

.zip of your game in web format: three_in_a_rogue_web.zip (will only work on Chrome when uploaded to any webserver, or served locally)
Link to web port: GitHub

That's so cool of you! My biggest complaint about love2d (because I really do adore this engine) is that it doesn't export to the web. I've heard rumblings of love.js and have been meaning to check it out but you got me a zip and everything! Thanks! I'll play around with the fullscreen thing to see if I can get it more natural on the web.

[NotExplosive]

Looks good! It took me a little while to understand how the match-3 mechanics worked and that they only happen when there are enemies around to defeat, but then the game became a lot of fun! Well done!

A few extra thoughts for improvement...
- As far as I can tell, there's not a lot of difference between X and Z attack when the level is empty?
- I didn't play long enough to discover if there are levels or if its truly just endless? It might be nice to have some sort of level structure and a minimap perhaps?
- Or another way to incentivise keeping playing would be a way to power-up your abilities or gain extra abilities?
- You might be happy with the randomness of the match-3 mechanic but if we knew at least the next row that was going to fall it would allow planning a bit more for strategy. Otherwise when you match a column you just have no idea what moves will follow!

Thanks for playing and thanks for the feedback. You've hit on a lot of things I plan on working on.

It is truly endless, however it's on my roadmap to create an actual handmade "dungeon" that has actual pacing and a difficulty curve rather than picking rooms at random which is what it currently does.

The randomness is definitely something I struggle with, currently it's possible to get a pure match from entirely new gems you couldn't have predicted would spawn there. That's normally is a "feature" of the match 3 Skinner box, but in this case it's actually really bad. I'm trying to reconcile if this is a game design problem (make random choices more informed) or a technical problem (prevent certain random scenarios from happening). I've already had to play with the weight of the randomness to make the "walk" gems common enough.

edit: Sorry for double posting, I'm still learning how to forum

[Davidobot]

That's so cool of you! My biggest complaint about love2d (because I really do adore this engine) is that it doesn't export to the web. I've heard rumblings of love.js and have been meaning to check it out but you got me a zip and everything! Thanks! I'll play around with the fullscreen thing to see if I can get it more natural on the web.

Fullscreen-wise, you can just avoid setting fullscreen/screen size based on screen dimensions. ie just set some predefined width/height in config.lua and then adapt your code to resize properly on love.resize calls.

It definitely is my biggest complaint as well. If LÖVE (semi-)officially supported web exporting, I think it will be one of the best game jam/prototyping frameworks. This is mostly down to the absurd amount of high-quality libraries.

The love.js port isn't perfect (mostly my fault, not Tanner's...) in the sense that games that use certain shaders (didn't have time to look into it) don't launch properly. Some other games need you to reload the browser once after a full load to work properly. Due to threading support, the port also only works on the latest Chrome versions, because of Firefox disabling shared thread memory due to (old) security concerns.

I should probably release a "compatibility" version that doesn't have threads but supports more browsers.

P.S. you shouldn't double-post, just edit the first post