So it's LD50! That's... a lot. My first Ludum Dare was #16 (yes, I had to check) waaaaay back in December 2009 (yup, got that from Wikipedia) and looking back at the archive website, it looks like I did pretty well too, with #12 in Fun! Sure, there were "only" 121 participants but it wasn't a bad start.

Funny thing, I actually ended up making the game in just around four hours. My original, much more ambitious idea involved building a level from tiles similar to Carcassonne, but I never figured out what the game would be, exactly. On Sunday evening, I just scratched that and quickly whipped up a game based on a very old (mid-90s) DOS game I had started but never finished, where you explore a maze while your torch slowly dies down. I had no time to polish it and it just used Unity's primitives and didn't have any sounds, but it managed (accidentally, I might add) to get a couple of things right:

• It was super simple to learn
• It was short

Those two are important pillars for LD success, I feel. Even though I did got them right at first, I did not do so well on all the the subsequent tries. All in all, from 2010 to 2014, I only managed to finish two games out of 8 I started. Out of those two, one was kind of a Civilization clone with an evolution theme (LD24), which was pretty ambitious and kinda worked out ok too.

From 2015 on I started a working in a team. That helped me make games that actually look good and also frees a lot of time, since even making my crappy coder art is very slow. Out of the Jam games, two have actually even ended up as getting published outside of LD: first Super Snowball Roller Mayhem 2000 in LD34, and second Ignis Universia, for which I'm actually working on a sequel right now!

For several years now, I've strived to play and review as many games as possible, with 100 as my yet unattainable goal. I think my personal best is around 70? The main reason for that being that I obviously like getting feedback myself, but there are always some gems to be found. To be able to play as many games as possible, I really, really, prefer playing web games for several reasons:

• It's the fastest way, as I don't need to download, unzip, etc.
• I'm antsy about running random stuff downloaded from the net
• They run on all (desktop) platforms

Last point is a bit moot now since I'm on a Windows PC, but I used a Mac exclusively for years, which prevented me from playing anything that wasn't native or web. So, if you're using Unity or other engine that supports WebGL exports, always make a WebGL version. There's really no excuse. Does your built-in-a-weekend jam game really need a feature not available in WebGL? Yeah, I thought so too.

Also, I still think we should have a story category!

Since everyone is sharing tips now, I decided to add some of my own. I have finished LD around a dozen times, most of that with a team and tried a lot of different things, even a multiplayer shooter once (did not end up well, don't try at home, kids)!

Preparation

Coming into the jam prepared can save several hours of time and, most importantly, nerves. Most of these are pretty obvious, but a good check list always helps

• Update all necessary software (including rebooting Windows)
• If using, setup a Git repo and share an empty project with all team members
• Install any plugins or assets you're going to use
• Do a test build, upload it and have someone test it
• Learn or re-learn workflows for the most important apps
• Check a few places for assets, even just for placeholder ones

Concept & Planning

First off, check out these tips.

Second, a thing you can see a lot of people doing here is coming up for an idea for each theme. That takes some off the actual jam and gets you going faster. Also, you can use the same ideas for multiple themes.

Deciding a genre beforehand

This is somewhat depending on how you approach the jam: if you really want to go in theme-first, it will mostly dictate the genre, but if you consider the theme more of an inspiration or a guideline, it can be useful to decide upon a theme or even a mechanic or two beforehand.

The advantage of this approach is that regardless of the theme (and as someone commented on the post linked above: the theme always sucks), you'll end up with a game. And that is the main purpose of Ludum Dare anyway. I also find it quite motivating.

Depending on your engine of choice, of course, you may need to do this anyway. Twine or RPGMaker lend themselves best to certain types of games and even with general engines like Unity make certain things easier than others.

Platform

My platform of choice has always (my first LD was in 2009) been Web, which mostly now means WebGL for games. If you want as much feedback as possible, it's the best choice.

• Multiplatform without extra builds and hassle
• Safer than downloading random games
• Fastest to get playing

The best place for distributing web builds is Itch.io, as its no hassle, no ads and reliable. Additional things to consider:

• Check that your game fits the window without need for scrolling at least on 1920x1080 resolution. I usually use 1024x576 but a bit larger window sizes should fit too.
• Have an option to turn sounds off.

Unity Tips

Here's a bunch of small tips for those of you using Unity: * 3D real-time is the most native way to do things. Sure, there's plenty of support for 2D too and making a turn-based game just needs some extra scripting, but you're also giving up things like NavMesh agents. * Make use of 1st/3rd Person Controller (pre-install!) * Unity Events are hand for level scripting, make a few general scripts that use them * Choose a font and at least placeholder UI assets early on so you don't need to waste time later to replace them * Especially important to test WebGL build before the jam, since there can be hiccups

Misc Tips

Start From the End

I know this sounds like a theme, but it's something I've been thinking about recently. Basically, when you test the game, the further you go, the more you have to play through the early content, or throw in some debugging options to skip that, which takes a lot of time that adds up during the weekend. So, if you make the ending first, and build back from that, you'll only need to test the new content (except occasionally, of course).

Onboarding

When judging, people play a lot of games and seldom have time to commit to a single one more than a few minutes, which makes player onboarding all the more important. Don't just throw all the info to the player in a single image or wall of text in the beginning, make it available as it becomes relevant, or, at the very least, have an option to review the how to play and controls.

Keep It Short

As said above, people don't have time to dedicate even an hour for every game, so make sure your game delivers its main selling points within 10-15 minutes. If it's a story-driven game, the whole story should be playable in this time.

In Conclusion

There's plenty of more I could prattle on about, but this post is already getting quite long!

In any case, the most important thing is to have fun! Making a complete game, even a small one, is an incredibly rewarding experience and the community here is quite amazing, so you'll get plenty of good and encouraging feedback.

See you at the finish line!

PLAY THE GAME HERE!

I know it's a huge to complain about the theme, but for me personally this was kind of a tough one. There are plenty of genres (tower defense, wave shooters, even puzzlers like Tetris and Bejeweled) and mechanics that fit right in to the 'delay the inevitable' mold, but they are all ones that I don't care much about. I've never been a competitive or achievement gamer, so racking up high scores while delaying the inevitable demise hold little glamour for me. In retrospect, having played a bunch of games by others, it's clear that many others didn't want to take that easy route either, and I'm seeing a wider variety of genres and approaches than usual. I think normally there are a few main routes people take in regards to the genre and a a lot of games are similar in that sense, but with Delay the Inevitable, at least so far, I'm not seeing that.

There's obviously a bunch of similar ideas floating around and one we took - flipping the assumption of the "inevitable" being a negative thing on its head - had been in some other games too (like this very crabby procrastination game). We started off with a mechanic that was traditionally pretty fitting to the theme, tower defense, mixed with having a controllable main character exploring the area. The first day was spent mostly making the essentials and putting stuff together, but on Saturday evening I felt like what we had was a pretty bad tower defense game, mechanically. For a TD game, having to do stuff with the character felt clunky and the menu I had built felt way too unwieldy to use, and just distracted from the action of managing the towers. In art and music we were definitely going to the right direction, with the game already looking and sounding pretty great. I also realised that since I don't really play TD games, I have little previous knowledge on how to design a good one, especially in such a short time! Around Saturday evening, I added a more narrative mechanic: short "prayers" to the petals we were using as a resource for building the towers. The idea was that you'd need to balance between fulfilling and sacrificing the prayers, depending on their contents. However, this just added to the overall hectic feeling.

So on Sunday ~~morning~~ afternoon, I backpedaled a bit and added the unlit towers, so instead of using the menu and cursor to to build them, you'd just light the ones already in the map. I also wrote a short intro to add to a more fairytale-like feel and a dialogue with the Demon Lord. This made the game a lot less about tower defense and more about exploration and story, which I think was ultimately the right choice, though one we maybe should have taken much earlier. Monday was used mostly for fixing things, finalising the map and balancing it more towards a slower story-driven game and the earlier fight against the inevitable horde of demons. I think, in the end, we did manage to get it quite right, though of course there's plenty that could still be improved. One thing is having better indication about the demons whereabouts, such as a minimap, or at least a bit more warning when they are approaching the Sakura tree. The feedback has been very positive however, so I'm ultimately happy with the result.

For me, this was a very learning experience in a couple of ways. One, I think in the future I need to put what I feel comfortable making above the theme, especially design-wise. We jumped into production a bit too quickly, not properly considering if the ideas we had were actually fun. That is, of course, how prototyping often works, but with LD that might mean a bit pivot one third or even halfway into the jam, when you finally start to actually feel the game. Secondly, we went with Unity's 2D and Tilemap and though I have some previous experience on them, there was a lot of new stuff I was not even aware of. Especially the Sprite Masks a very useful tool I'll certainly make use of in future 2D projects. There are limitations, of course (e.g. you can determine from script if a point is inside a mask or not), but overall it was quite exciting to get to use the new 2D features.

I had a lot of fun working with @cheezegami and @danae and I think we had a good shared understanding on what kind of a game we'd like, even if getting there took a bit of time for me.

I've done so many LDs I've lost count, but this was the first time I actually finished the Theme Slaughter! I don't know if there were a bit less themes than usual because of the new 2 suggestion limit, or if it was because I've already decided I'm going to make an RPG for LD51, so it boiled down to "can I make an RPG out of this?" (ok, there were other things like not keeping themes that have been used previously) and thus it was a simpler process. Also, a lot more keeps than I have usually had (it's been like 1-2% previously).

So, yeah, I'm in and whatever the theme is, my game is going to be as full-blown RPG as I have time for. Hoping to get at least character creation, items and level progression in, with around 20-30 minute story and some combat.

I mean, what else do you need?

Over the years I’ve had plenty of assumptions on what kind of games tend to do well on Ludum Dare, but I’ve never actually really looked. So, with a bit of free time over the holiday, I took a quick glance at the 10 most recent Jam category winners. The sample size is small and there are not many conclusions that can be drawn, I did find some interesting things.

Edit: Just as a disclaimer: I'm not in any way suggesting that you should try to mimic these games, or that trying to win is the best way to participate (it isn't). The sample is also way too small to say too much about what kind of games tend to win (to do that you'd have to probably take in at least top 10 and try to find some larger trends)

First Off: Team size doesn’t correlate with rating success

This is maybe the most counterintuitive one and I would have assumed the best rated Jam games to at least have a couple of members per team, but in truth half of them are by solo developers. I didn’t look into reasons why they have been submitted to the Jam instead of Compo, but I would guess it’s mostly to either get one extra day of time, or to be able to use third party music. Only one game had what I’d call a large team, that of The Not Very Golden Age of Piracy with 7 members.

Second, most winners have an average number of reviews

There was one very clear outlier here: The Shotgun King with whopping 243 reviews, but most get around 30-60 reviews, which I would assume is a pretty average number for any Ludum Dare game (I don’t know if there’s data of this anywhere). Some have voiced complaints that game with review numbers barely above the required threshold can game the system, but I don’t see much evidence of that here.

Third: puzzle games seem to perform pretty well

I’m not sure if it’s because they are relatively simple to make and thus leave more time for polish, or if that nice feeling you get for solving a riddle just makes people rate higher on average, but 6 out of the 10 games could be categorised as puzzles. There was plenty of variance within that category, but with platformers being another popular choice, two of the winners were a combination of those two.

Most of those puzzle games were still rather arcade-y, and only 3 games did not contain any action or coordination elements.

Fourth point is the about the visuals

If you thought most games in Ludum Dare utilise pixel art, you are probably correct, corroborated by the winners. Only two of the games were 3D, and even on of them (Dark Points) was low res in presentation. I think this comes down to pixel art being relatively fast to produce with a good quality, and a big portion of participants falling into an age range that has either fond memories of pixel art or is active in the indie games scene where the artstyle is still popular. It should be noted that two of the games fall in the subcategory of NES pixel art, with Burning Ravager being especially faithful to the old console’s limitations.

Finally, familiarity with a twist seems like the winning concept

Most of the games relied either on established genres or popular games as their basis and usually added a small twist in the gameplay. Most of the games needed little or no tutorials, and got you to the ‘beef’ of the gameplay very quickly. The notable exception here is Fallumns, which you need to play for a few minutes for it to reveal it’s hand, but the twist was clearly good enough to elicit high ratings for anyone who sacrificed a few minutes of their precious rating time. It doesn’t seem like the length of the game had much effect on the final scores, as the playtimes ranged from around 10 minutes (Lost in Translation) to practically unlimited play (Shotgun King) with no trend towards either end.

Some final thoughts:

• None of the games were especially story-driven. Not if stories are just too divisive to get an all-around high rating, or tend to make the games a bit too unwieldy
• There’s a strong correlation with high overall score and the rating in all other categories (with the possible exception of Humour, which was not applicable to all the games)
• Sounds and music were mostly 8-bit/chiptune, which was to be expected. The voiceover of The Not Very Golden Age of Piracy made it standout and contributed to a very polished feel.

I know this is hard to believe for many, but for once, I think that the theme didn't suck. While I don't think it was the best option out of the last round of themes, it was a solid one that left a lot of options while not being too broad. Sure, many of the games I played did fall under some kind of farming category – and we had those elements, too – but they were quite enjoyable for the most part, unlike with 10 seconds, where it felt like the majority of them would have been better without the theme tacked on.

We'd discussed making a Zelda-like and, after the theme was announced, quickly settled with the idea of combining that sort of gameplay with growing your own weapons. The latter mechanic premeditated having to change your weapon regularly, so we ended up with the short durability of the Breath of the Wild and a general look-and-feel closer to the NES and SNES Zeldas.

Overall, I feel the jam went quite smoothly this time, and we didn't end up making any drastic changes in direction or big gameplay overhauls. As usual, a lot of the details crystallised during the weekend as the game started to materialise, and the story the game should have had only afterwards (which is why there'll be one last big post-jam update soon). We had a relatively large team but didn't hit many snags with communication or organisation. One thing I wish we'd maybe planned better was the enemies, as I didn't see the scale from the early concept art drafts, and we ended up having too large, almost boss-like monsters (in addition to the boss itself!) and only one normal-sized one. I was able to address the unbalance in the enemies by adding one that was visually just a simple particle effect, though, so I don't feel like it was a bad problem.

Once again, I'm super happy with how the game turned out visually. The art certainly stands out not just in quality but also in style, with the quirky use of colour and the, uhm, happy trees (RunicPixels wrote an article about what went into the tilemap here).

(Concept art by Rumboots)

Technically, we made the call before the jam not to do anything too experimental this time (isometric maps ended up taking much more time in the last jam than I originally anticipated), so I was mostly just implementing things I already knew how to do and tweaking the design. One thing that did cause more problems than I thought were the cameras, as I wasn't aware you can't constrain a Cinemachine camera to a simple BoxCollider2D. It wasn't too much work to roll my own constraint system, but it's still a bit wonky.

It was also a great relief on Monday when, in addition to again composing the amazing music and doing the sound effects, Danae stepped in to help and coded a lot of the small polishes that got the game across the finish line in much better condition.

I'm also happy that I had time to add a simple tutorial/onboarding to the beginning of the game, as it looks like it helped a lot of people to grasp the gameplay a bit better, and it's something that I usually don't have time for. In the future, I'm planning to add some even earlier, though, and make the system a bit more robust. It probably says something that I actually have the energy to make one big update now that the jam is over, and it won't mess up the rating, overhauling the story to the idea I got a bit too late and adding a few extra touches based on the feedback we received. I'm not convinced Warden of the Woods is a game make an actual release out of, like Ignis Universia was, but it ended up as a good game for Ludum Dare, the purpose it was made for.

rolls sleeves

It is done!

I don't have that much to say about the themes, they felt to me to be same as usual, though I think I generally end up flagging more than one. I still can't understand people who suggest themes that are the same as even recent ones, what's the point?

Collapse: Shoot weird stuff at wizard tower to collapse it. Scorched Earth-like maybe?

Rock, Paper, Scissor: Fighting game with 3D characters and cool anims, lots of juice, but ultimately it's just RPS

Alchemy: Make potions, some combinations explose

Strange Creature: open world game, find and pet all creatures

Edge of the World: You're a giant creature who has to balance on a flat earth

Repair: Recettear but it's a repair shop and you have to go to a dungeon to get spare parts

Build your own path: Minecraft-like where you need to build a path over a terrain with crafting

Single Resource: You're a vampire and your blood is both HP and MP (also, my suggestion, yay!)

Momentum: Snake game?

Leave something behind: ... dungeon crawler but after each level you need to leave one item?

Connected: Social deduction game, who knows who, like Murdle

Combinations: Arrange people into teams to solve puzzles

Isolated: Escape room but bigger?

Stronger together: Tile-based strategy game where allies in adjacent spaces give bonuses

Sanctuary: Cross a wasteland that's has areas where you can rest and regain health

Folklore: Card game with different folklore creatures

Keep the light on: Explore a maze but your torch keeps running out, find torch refills. This was actually my first LD back in the day then the theme was:

Explore: Find collectibles in a proc gen forest using different methods of traversing.

Castle: Rampart x tower defence.

Keep it contained: Lemmings-style game but keep the ones who know THE THING from leaving.

Roots: Forest sim, grow roots to find water and resources.

Treasure hunt: Deduction game/puzzle/exploration: you get a map with weird symbols and clues and have to find the treasure in a large(ish) terrain.

Piracy: Bring a pirated copy of Sid Meier's Pirates to a pirate with a pirated time machine.

Sacrifice: You're an evil cult leader but also a vegan and have to figure out substitutes for the sacrifices in rituals.

Garden: Stealth/heist game where you steal veggies from a garden as a bunny/mole/whatever.

Single use: Dating game, but if you use the same pick up line twice and the datees find out, you're screwed.

Power source: NPC gives you awesome magic powers, but also goes faster than your walk and slower than your run.

It spreads: Use mouse to spread butter on a bread without spilling. The shapes of breads get increasingly more ridiculous.

Strength in numbers: Strategy/puzzle: your troops in each tile double every turn, but if the capacity is reached you lose. They can spread to adjacent tiles, but only if the amount of troops is divisible by the # of free tiles.

Strange machines: Explore a strange world of machines and figure out what they like to lure them to their enclosure.

Mutation: Simple arcade game (e.g. shooter) but every power-up changes the controls randomly.

Summoning: Deduction/exploration game where you need to find out demons' favourite foods to summon them.

I have to admit I struggled with these a lot more than the previous two rounds. Also difficult to not reuse some ideas since the themes are so similar!

Cartography: Text-based dungeon crawling but there's no automap and you have to draw the map yourself.

You don't play as the main character: Sherlock Holmes, but he's not actually very bright and you play as Watson and have to arrange the clues in the right order for him to pick.

Recycle: Scavenge the pieces to build your own ultimate CYCLE OF VIOLENCE (https://www.smbc-comics.com/comic/a-vicious-cycle).

Tunnels: First person, you see a map of the target facility and have to burrow under it, tunnel length is limited and you can also find bonuses if you're clever.

Pick your poison: You play as a dyslexic French fish salesman. Or something related to the Affair of the Poisons, idk.

Decay: The Grog Bartending Simulator.

Digging: You command a colony of dwarves and dig a 2D sideview underground city, but have to avoid Balrogs and other monsters and find gold.

Fishing: You're the worm and have to steer the fishing line to avoid getting eaten.

Push your luck: Build a tower that becomes more and more unstable and the parts that give more points are the hardest to fit.

Out of reach: Traversal-based puzzle where you see the goal in the beginning and have to figure out how to get there.

Symbiosis: Puzzle game where organisms consume and produce different chemicals and you need to find the right balance.

Shelter: Foraging RTS focused on the resource collecting side of the genre; forage resources and build shelters, but when the acid rain starts falling, the foragers only have limited time to reach the shelters.

Clockwork: Time sensitive puzzle platformer where you have to learn how the (one screen, let's not get crazy here) machine works to reach the goal.

Reflection: Metroidvania where you use mirrors to enter a mirror-world that is mostly identical, but certain objects have been flipped on the x-axis.

Out of order: Puzzle game where you have to order objects according to obscure criteria (albums by the age of the cover artist, games by the number of letters in their engine name). The criteria is not divulged but it shows which two objects are in correct order to begin with. Googling is encouraged.

Change the rules: Roguelike but the character is automated and you have to alter the generation rules for it to reach it the end of the maze.

So I decided to make another idea for each of the final round themes! I tried to be a bit more realistic here, though.

Summoning: Cozy game where you summon little demons to help with daily tasks.

It spreads: You're an alien pod people boss and have to manage the invasion: order minions to replace others, you have to separate them and clone them in secret.

Recycle: Collect the healing potions and other consumables left by the adventurers and upcycle them and sell them in your shop.

Tunnels: First person tile-based game where you explore a tunnel network and have to find places that correspond to the overworld above.

Strange machines: Machines litter the landscape, figure out what their parts do and switch them to complete puzzles.

Single use: Action RPG but the weapons break after one hit. However, you auto-equip the next item from the inventory, so inventory management becomes key to survival.

Strength in numbers: Strategy game where you build buildings and the strength value of each building is the combined value of its neigbours. The values can also be negative, or multiplications or divisions. Each new building is chosen for a small randomly selected set.

Fishing: 3rd person arcade action: play as an orca and take down as many fishing boats as you can. The fishers start using increasingly aggressive weapons as the game progresses.

Push your luck: Simple dungeon crawler, room by room. After each encounter, you can decide to either go home or continue onwards. Rewards increase exponentially, but your resources dwindle.

Digging: Graveyard manager, dig graves for recently deceased, but you have to consider how they rise up, zombies and vampires just don't mesh that well.

Cartography: You have a map marked with some hotspots. Explore the locations and choose the correct explanation for the place.

Pick your poison: Very simple roguelite gameplay with short sessions, but before each run you pick an item that has both an upside and a downside. Could be bar-themed.

Decay: The city decays before your eyes. Bridges and walkways disappear. Collect cultural relics before paths to them disappear.

Collapse: 2D platformer where you jump on platform to collapse them on enemies.

Mutation: You're a monster farmer and clients come to you with different requests. Combine random mutations to complete orders.

You don't play as the main character: Escort quest where you are the one being escorted.

This was the 20th time I successfully created a Ludum Dare game!

Since it's 2 AM in here and I have to wake up in 5 hours to go teach some game programming, I'm going to reminiscence later!

First off, play it here, if you haven't already!

Ok, so the game is too hard. I know. Funny thing is, this is pretty much the first time I actually had a testable build early enogh to get some feedback, even tough I had less time than usual since I only started on it very late on Saturday. But unfortunately, most people didn't have time to test it so I was left with very few data points. The original version was, unbelievably, even more obtuse than the final one.

So the main thing that went wrong was definitely the difficulty. It was a hard balance to strike with having the right feel for the game, but still making it playable, and it erred too much on the side of obscure. I also wish that have had a bit more story, and maybe some other characters in the game. And should have recorded some voices for the demons.

Few comments wished the summoning cutscene to be skippable, but that's actually one design decision I stand by, since the longer summon sequence deters from brute forcing the puzzle.

Art-wise, I think I managed to nail the style pretty well, though I'm not 100% happy with all the symbols. Judging from the comments, people generally appreciated the visuals. And the music, which comes as a bit of a surprise since it's the first "song" I've composed in over 15 years and, honestly, I wasn't very good at it back then either. I played it safe by googling chords suitable for a horror soundtrack and then just building it around that.

Overall, I'm happy with the result, even if I wish I would have been able to test it more and tune the difficulty downwards. On to the next 20 Ludum Dares, then!