lochi.jpeg

Rating systems are cooked

lochi — Tue, 07 Apr 2026 24:00:00 GMT

Rating systems are cooked.

But I still use them.

Comparing pieces of art based on a numbered rating system is inherently irrational because obviously not every work in a medium can be directly and objectively compared in quality.

For example, the aim of an MMO like World Of Warcraft is going to be a lot different to a visual novel like Ace Attorney, which is going to be a lot different from an experimental indie game like Yume Nikki. Judging both on the same rating system and saying one is better than the other is comparing apples and oranges. Different target audiences, different goals, different amounts of time that need to be committed.

Now that hasn't stopped me using sites like Goodreads, Letterboxd and Backloggd to log the media I enjoy because it's still genuinely fun to rate things so you can track what you like and then catalogue that data for easy access.

Score aggregation

On these review aggregation sites like Goodreads, Letterboxd and Metacritic everyone is bound to have their own system for how they rate media. Given how arbitrary the system is already the only real guarantee is that higher is good and lower is bad.

My reviews on backloggd are centered around 4/5

When put under these circumstances, I've noticed people tend to rate everything with a positive skew. I mean it makes sense. Why would you watch something if you aren't going to enjoy it? If you look at any rotten tomatoes, almost every film has a much higher audience score than critic score.

This is the first movies on the rotten tomatoes home page at time of writing. Every film has a higher audience score.

As it turns out people reading critic reviews also tend to hope that said critic likes what they like. That's why it is also interesting when people attack reviewers for giving something a "low score". 7/10 might mean middling to you, but to the critic that could be a high score on their own scale.

This is why I think that for someone to make a rating, there should be some form of criteria that the platform abides by in order to determine what each level of praise really equates to. It would help level out the distribution of ratings whilst also hopeful making ratings more unanimous among the population if everyone plays by the same rules.

I also feel that more major platforms should omit numerical scores from their reviews, letting the nuance of their writing speak for itself. That won't land them on Metacritic though I suppose.

The urge to have the correct opinion

I don't know if this is an urge that only I have, but often I find myself looking at rating sites almost religiously, whenever I plan on watching anything at all. I mean how do I know if it's worth my time? It's such an irrational urge, but I find myself coming back to it all of the time. It almost saps the fun out of everything when I've read 5 reviews of something before even sitting down to start. This is also tied to an inclination to avoid trying new things. It's much more comfortable sitting back and reading reviews rather than going through all of the effort to commit to something new.

Also, when I watch something and am not entirely affected by it, despite a high public rating, I often find myself working to discover what made it good to everyone else. Sometimes this deepens my appreciation, like with Outer Wilds, and sometimes it makes me realise just how arbitrary these ratings really are. Like with Mario Galaxy. (I do not like Mario Galaxy. Yes. I am talking about the game)

Then again it's not all bad. I've found heaps of my favourite music through sites like rateyourmusic and I've enjoyed plenty of movies that I found through Letterboxd. Maybe I'm just concerned that I am not striking a balance. Without seeing something bad is it possible to appreciate quality and effort? My answer, at least for games is by making my own. Once you know how games are made, all of work put into making what I play is able to shine through clearer. For films though? How do I know if one is special, when I force everything to be special.

Ditch the aggregator

If I were to become a wise old man, I would ditch the public review aggregator entirely, go on my private notes and write down reviews for every single piece of media I experience. Unaffected by the masses, the aggregated, indistinguished average opinion. Well I'm going to try that out now.

There was recently a change of community guidelines on backloggd that I didn't like the sound of. Essentially they gave more free reign for moderators to abuse their power and eliminate any review on the platform that doesn't stay precisely on topic. I don't really care too much about this, but I was interested in leaving anyways to achieve my very important ambition of becoming a wise old man.

I downloaded, processed and exported all of my game logs into individual markdown files alongside their cover images. I then created an Obsidian vault dedicated entirely to tracking any media I am playing/reading/watching, with queries and tags to categorise them into past, present and future. From the compressed images sourced through Backloggd, the images for 237 games totals up to 5mb of disk space, which is pretty modest for how much it adds to visualising the data in a grid.

This is a visualisation of my game notes, using obsidian bases. Each file is an individual note.

This is the note for one game.

I might merge this system into my primary vault. Soon I will also remove backloggd links from this website. Maybe a recommendations page will be nice, to show a bunch of game recommendations alongside some briefer reviews of games I played in the past. Who know? I might try reviewing some other mediums, although I'm hardly qualified in judging much outside of games.

Bonus for reading until the end: Rating site ratings

Rotten tomatoes: Rotten. I don't like it. Never feels accurate to anything.

Metacritic: Pretty good. That is until slop live reports on how Game of Thrones fans are review bombing Breaking Bad. Rating drama cannot be this deep. Partially responsible for pushing reviewers to take on a numerical rating system so they can appear on their page.

Rate Your Music: Good, if you ignore certain individuals. I like this site for finding super niche music recommendations. They also catalogue complete artist discographies, including bootlegs so you can find some super obscure stuff.

Backloggd: Alright for finding games. Guidelines are worrying, a little pretentious. As if I can't pretend I don't come across a little prentious in a lot of my blogs lol.

Letterboxd: You either find the most pretentious "cinephiles" ever, or normies. No inbetween.

Goodreads: Laggy to the point of being almost unusable. How do you mess up this badly? I don't like reading book reviews much either.

Flip and fall sequel - Devlog 2

lochi — Sat, 03 Jan 2026 24:00:00 GMT

Its been a little over a month since my last devlog, I am still hard at work on this game.

Lighting

So the first thing I worked on since the last devlog was the lighting system I'm implementing lighting by hand. Its done through a lightmap texture thats rendered at runtime based on the lights in the scene, that then gets multiplied by the unlit scene. For each light I block of the geometry and add it to a per-light shadow map texture.

When each light is then rendered to the lightmap texture, its cut out around the level geometry, making some neat looking shadows. So much for prioritising important work, but I'm pretty happy with the result. This is heavily based off again one of Noel Berry's blog posts, as well as some extra reading from a commentor on the post.

Broken lighting that looked cool.

Correct lighting.

Player movement

The player's movement has gotten a lot of work in this time. I've added a momentum mechanic, which will store the players momentum when jumping off of moving obstacles. They also have a dive move, that you can then jump out of.

The dive can also be performed on the ground to slide underneath tight spaces.

I want the dive to sort of have a risk factor when using it so it leave the player with a lot more velocity than a regular jump, meaning that if you aren't careful you can slide off platforms after landing.

Later, I might change the dive to look a little different in the air than on the ground.

Time travel

Time travel in this game work through portals that transport you to and from the present. Portals transport you to the past version of a level. For example, the first stage has you exploring the abandoned android factory. After passing through the portal, you will get to see what the factory looked like back in its prime.

You will also be equipped with a handheld time travel tool that lets you rewind the screen to its original state. You will have to use it to solve puzzles and collect all of the secrets left in the world.

Level gimmicks

So I also built a few objects that you will be able to interact with in the game. These gimmicks will completely change based on what time period you are in.

This is a conveyor belt in the present, due to inactivity in the factory, exposed wires have made touching them unsafe.

Back in the past though, the conveyor belts are working perfectly and if you time your jumps carefully you can use its speed to launch yourself even further.

Combat?

I might experiment with some sort of combat because the game obviously has androids as the enemies and it would be odd to not fight them at all. I really want any combat to weave well with the heavy focus on platforming because a lot of games really separate the two systems when integrating them together could add extra depth to both systems at once.

Take for instance sonics homing attack, that turns combat into an action that also serves as a platforming mechanic, using it for both attacking enemies and for quickly navigating to smaller targets like bounce pads.

Image credit: Sonic fandom

Wrap up

Whenever I talk about my games in person a lot of people ask what it's called. I usually have no idea. That's all I have to share for this devlog!

Flip and fall sequel - Devlog 1

lochi — Tue, 02 Dec 2025 24:00:00 GMT

Programming in FNA/monogame feels really alien to me. You really need to plan your code out a lot more meticulous than you would have with Godot, because everything down to managing what you are drawing and updating is up to you to implement.

I'm using LDTK to create my levels, with it's JSON parser to load them into my game. The whole level editor format is stored as one big json file which is super convenient.

I also started by using monogame aseprite to automatically convert aseprite files into loaded assets. This worked fine but I wanted to minimise the projects dependencies.

Later, I replaced this with my own solution, a script that I run at the start of debugging that uses the aseprite cli to compiles the a spritesheet image and outputs a json file for the metadata to group sprites into tags.

aseprite -b $file --list-tags --sheet AsepriteOut/$BASE.png > AsepriteOut/$BASE.json

aseprite -b flag turns off opening gui
file.aseprite --list-tags includes tag data in final json
--sheet outputs a spritesheet
AsepriteOut/filename.png spritesheet path
> AsepriteOut/filename.json Redirects the json which is printed out to the console into a new json file

This means when I export the final game, the aseprite files won't be shipped, just the spritesheet png and it's json metadata.

Using the new asprite importer, I started making some character animations. Now you might notice the character is bald.

This is because I made some hair physics, which is a chain of circles that fall downwards, and are pushed around when the player moves.

The run combined with the hair physics ends up looking like this!

I feel like an imposter given that this game copies a lot from Celeste. As I continue developing, I'm going to try and make a conscious effort to separate myself from inspirations.

A jam game, and possibly a sequel - Devlog 0

lochi — Fri, 14 Nov 2025 24:00:00 GMT

So a couple of weeks ago now I joined the Untitled Game Jam #117, and made the game Flip and Fall. It's a four level long 2D precision platformer made in Godot, heavily inspired by Celeste.

Since then, I've been stuck in a bit of a limbo, trying to learn a new game framework whilst having no clue what kind of game I want to make.

After a couple of days of messing around and making random stuff, I've finally settled an idea on what kind of game I want to make.

So recently I read 'Do Androids Dream of Electric Sheep', the inspiration for the film Blade Runner. After I finished reading, I figured the structure of the plot seemed pretty fit to adapt to a game. I won't spoil the plot, but the basic premise is Rick Deckhard, a bounty hunter is assigned to hunt down a group of rogue androids. Now this sort of fits in with the setting of Flip and Fall, which was set in some sort of vague factory setting with mechanical animals.

So the plan for this new game is a sort of reimagining or fleshing out of the original flip and fall, maybe with a proper narrative longer campaign and better art.

I'm developing this game using the FNA framework, inspired by the blog post Noel Berry - a developer on Celeste- on making games without using engines. I've only ever used Godot to make games so I figured it would be a fun challenge to learn C# and build a game with less pre-made tools on hand.

Just throwing this out as an idea, but I also want to experiment a little with music production, specifically ambient maybe in the style of Radio Amor - Tim Hecker or Richard D James Album - Aphex Twin.

As with any game I start this is subject to never be finished at all, but hopefully this devlog series keeps my drive to keep developing up.

Pikmin 3 and 4

lochi — Fri, 14 Nov 2025 24:00:00 GMT

Pikmin 3 Deluxe

Pikmin 3 is a lot like the original Pikmin. It's ditched the addition of caves from Pikmin 2 and returned to the formula of crash landing on a planet and recruiting the Pikmin to help you escape. The plot is a bit more convoluted than the first two games. There are now three protagonists who are on a mission to feed their starving home planet.

Over the course of the relatively short campaign, you unlock all three captains which can each be controlled separately like in Pikmin 2. However, now you're able to use the map and automatically route your captains to walk to specific locations, allowing for a level of optimisation not possible in Pikmin 2, where the only squad that could move is the one that you controlled.

The balance here also felt a lot better than other games. The rock Pikmin were the main standout as being overpowered, however the way you approach combat with them is a bit different to the other Pikmin, because they are unable to stick to them. This means to fight more efficiently with them, you can throw them yourself over and over again, requiring more of your attention. The winged Pikmin are also specialised, being good for a few certain enemy types and for carrying objects over obstacles.

Pikmin 4

I spent a lot more time with Pikmin 4 than I have with any of the other Pikmin games and I even went for full completion. It has by far the most content, spanning 5 more open stages, and a command post where you can interact with the various npcs.

The opening to this game is slow. You're greeted with an hour or so of mostly dialogue between you, the silent protagonist and the rescue corps. The dialogue only really slows down once you finally get in the thick of exploring the stages, which are much larger than in any previous game. The camera is set lower, making this feel more like an action game than an RTS, an idea that's further reinforced by the lack of a second captain.

The second captain here is reduced to the most overpowered mechanic in any Pikmin game thus far. Oatchi, the two legged dog that hops around behind you or can go off with his own party of Pikmin. He can do a million things including

Carrying your entire squad Pikmin on his back
Ramming into and stunning enemies
Swimming over water (with the squad on his back)
Lifting with the strength of a hundred Pikmin
Attacking enemies on his own

The only downside to Oatchi is that you can't pass through certain gates with him because he is too large, and you can really feel how much less powerful you are when he's missing.

The caves are also back, which I feel are much improved over Pikmin 2. They feel like the complete opposite of those caves because they feature no procedural generation and are at times a little too easy instead of too hard. The game did have much better assist features than the second game, letting you rewind back to the start of a sublevel to retry.

The good

The post-game is amazing. The dandori challenges really added another layer of organisational depth that you could only really find in some of the challenges from the third game.

They even retell Olimars crash landing in the lens of this alternate universe, recycling the familiar location from the base campaign and setting you out on another treasure hunt with a stricter time limit.

Nitpicks

The upgrade system feels superficial and like something they just HAD to do because it was a AAA game that needed a progression system rather than something that genuinely improves the gameplay. For an RTS like this, most of the upgrades that just boosted stats should have been left out in place of just the new abilities because by the end, I felt completely overpowered.

Conclusion

Overall I really liked Pikmin 3, it shakes up the original Pikmin formula just enough to stay interesting throughout its admittedly very short runtime. I also enjoyed Pikmin 4 for its larger scale and bigger worlds to explore, although I feel that with it's increased scope, it suffers from a certain amount of bloat that softened the challenge to cater for a larger target audience.

Where to listen to music but better

lochi — Tue, 15 Jul 2025 24:00:00 GMT

Editted 12/12/25

I listen to a lot of music. Pretty much every moment while I work I have some sort of music on in the background. Thus, I've spent a lot of energy choosing the searching for the best music player.

There is an abundance of music players to pick from including alternate clients for spotify and general music players for local files.

Spotify

The quickest and most convenient option for most people would to be to use either the spotify desktop app or website. One thing to note is that spotify doesn't actively support the development of the official linux client and but let engineers to do in their free time.

Spotify also isn't a very ethical company, most recently the ceo investing 600 million into a german military software company that manufactures strike drones. Great to see where the spotify premium money is going.

If you do still insist on using Spotify, Spicetify lets you modify your spotify client to add extensions themes and more. Extensions include visualisers, ui tweaks and adblock to avoid paying for premium.

Alternatively, there are various open source spotify clients however that offer slightly different features.

NCSpot

ncspot is an ncurses Spotify client written in Rust using librespot.

ncspot is a very simplistic alternate client that runs in the terminal. Compared to other options, its lacking in features but lets you do all of the essential functions of listening to music and searching. It's a good option if you want the least possible performance impact and aren't interested in discovery features or album art.

spotify-player

spotify_player is a much more featured spotify player than ncspot. Its got more decorated ui and has support for rending the album art.

Despite the various alternative options, I still use the default desktop app, all other clients are lacking in terms of discovery tools for finding new music compared to the recommendations of the normal app.

Downloading music

There are various legal ways you can get music from online onto your device.

Listening to albums for free on bandcamp
Purchasing music on bandcamp and getting high quality or lossless downloads
Download official from soundcloud.
Using a downloading tool to rip from youtube or soundcloud such as yt-dlp
Find artists that have free downloads like death grips (explicit album art)

Bandcamp

Bandcamp is very lenient with letting users listen to music for free. For most albums on the platform, you can listen to them for free multiple times before needing to purchase. When you do purchase an album you get complete download access to multiple formats.

yt-dlp

yt-dlp is a command line tool that allows you to download audio and video from various online sources. It's pretty useful if you find a video online and aren't able to download it normally. You can paste the link to the content into yt-dlp and it will download it for you. It's quicker and way less sketchy than those youtube to mp3 websites.

To download just the audio of a youtube video you can use this command in the terminal for yt-dlp

yt-dlp -x -f bestaudio URL

The -x flag extracts it to audio and -f bestaudio makes it download only the audio format data.

You can also use yt-dlp -t mp3 URL to download as an mp3 file.

The arch wiki lists out some more useful yt-dlp commands.

Music players

With your downloaded music, you'll obviously need a program to play it. These are a few of the options that I've tried thus far.

Windows

Foobar2000 and musicbee seem to be the most popular choices for windows music players right now. Musicbee is fairly simple to set up and has a very customisable UI while I've found foobar is for more technical users. Personally I like musicbee given its relatively modern and simple to setup.

Linux/Macos Terminal clients

cmus

cmus was the first music player I got working. It's lightweight and simple to set up. It has all of the necessary features to play your music and not much more. I was pretty content with using this for a while before I found rmpc.

MPV

Not to be confused for mpd, mpv is an application for playing media files and works perfectly fine for audio files as well.

MPD Clients

MPD is a popular backend for playing music, primarily on linux systems though it may work on macos and windows. There are dozens of frontends available.

It takes a bit of time to set up an mpd configuration but not many options need to be enabled in order to get it into a functional state for just playing back local files on a single machine.

One of the biggest draws of using MPD to me is that MPD itself is a daemon which means that the backend for your music player is competely seperate from where you control it. This means you can close your client and the music will still play in the background, saving you from needing to keep another window open in the background.

For a mostly complete list of clients see MPD's website but I'll list the ones I like below.

It is possible to still track your listening activity using services like last.fm on many of these players including mpd. I use mpdscribble. Last.fm is used on my website's homepage to show what I'm playing.

RMPC

RMPC (Rusty Music Player Client) is my favourite MPD client right now. It looks greats, is more customisable than any other player I've used and is still getting developed and improved right now. It displays album art amazingly and lets you customise exactly how big you want it to be. See this video by bread on penguins.

NCMPCPP

The worst named piece of software I've ever seen. The title stands for NCurses Music Player Client Plus Plus, inspired by another mpd client called ncmpc. Development for this project is currently inactive but its a decent minimal client.

Wrap up

Right now my current setup consists of mpd with the rmpc client. My configuration can be found on github if you are interested.

Pikmin 1 and 2

lochi — Sat, 12 Jul 2025 24:00:00 GMT

Pikmin 1

I really enjoyed Pikmin 1. You play as Olimar who crash lands on Earth, scattering the parts of his ship. You recruit and manage the pikmin to help you fight various insects and carry the parts to reassemble your ship. Theres a time limit of 30 in game days before Olimar runs out of oxygen and gets turned into a pikmin himself.

I played the game first a couple of months ago and quit midway because the time limit was too stressing. This time, I just made sure to get a ship part each day and was totally fine.

Managing all of the pikmin and exploring the scaled down planet was super fun. Keeping your squad alive and figuring out the weaknesses of enemies was satisfying.

The pikmins movement ai is a little fiddly and they get stuck on literally everything if you don't manage them carefully. I like it though because the pikmin are supposed to be a little dumb.

Overall, I really liked Pikmin 1. The soundtrack is relaxing and its simple enough to jump back in again and go for a shorter run.

Pikmin 2

I have a lot more to say about the Pikmin 2 because I didn't like it nearly as much as the first. In the second game, you return to Earth along with your co-worker Louie to collect junk to sell and repay your company's debt.

The bulk of Pikmin 2 is spent in caves, a departure from the first game which is entirely above ground. The caves are dungeons where you clear out each floor, grabbing all the treasure you can while losing as few pikmin as you can.

The new Pikmin types are the purple and white Pikmin. The purple are good for combat being able to do massive damage and stun enemies. The white are able to traverse poison and dig up hidden underground items.

The good

Having 2 captains is really useful, opening the way for more productivity and giving a nice risk/reward of having more productivity but more risk of danger because you can only control and see the view of one captain at a time.

There is also a lot more treasures here than in the first game, mostly product placement from real world brands. The switch remaster gets rid of all the branded products though.

The piklopedia is fun to look through and see Olimar's observations about all the treasures you collect.

The bad

My biggest complaints with the game are to do with the caves. Nintendo removed the quick save loading from Pikmin 1 that this game sorely needed. In the original, you could press a button and reload from a previous save. Pikmin 2 is a lot harder with a lot of one shot kill enemies that could wipe out your whole squad. To reload your save here you need to restart the whole game. It might be worth playing it through an emulator for the save states.

I really didn't like a lot of these caves either. It was a chore to get through the tougher ones, only because of the aforementioned save loading. I stopped at the start of the post game content because the difficulty just spiked up exponentially and it was become too much of a chore.

Time is also paused in the caves which encourages going through them a lot slower to save your pikmin unlike the original game which prioritised going at a fast pace and aiming for efficiency.

The caves are procedurally generated so in almost all of them, visuals are incredibly barebones compared to the overworld of the first game.

Focusing on caves and combat instead of the exploration and time management above ground shifts the tone of the game a lot to feel more like a dungeon crawler rather than an exploration game which in my opinion, makes it lose what made the original so fun. Ditching the core loop of gradually exploring the terrain to uncover treasure and managing time in favor of a slow, grind centered much more on combat.

All of my complaints can pretty much be boiled down to "skill issue" so maybe don't take this that seriously.

Conclusion

At some point I'll be playing Pikmin 3 and 4 I guess because I've heard that 3 is a lot closer to 1 and 4 is sort of a more fair version of 2. Right now I'm playing re4 remake.

Finishing Mecha Market

lochi — Mon, 12 May 2025 24:00:00 GMT

I finished Mecha Market. It started development on September 1st 2024. It's about building and running a grocery store, heavily inspired by roblox retail tycoon 2. It features a full layout editor that lets you completely customise the layout of your store.

The biggest change to my workflow for this game was using Obsidian to track my progress. Taking notes helps a lot with keeping track of what needs to be done, speeding up development.

According to tokei, I've written 5503 lines of code, excluding blank lines and external addons. My previous largest project, GDFP was 3455 which makes this game a 59% increase in code size.

A lot of parts from the game were taken from Brass Bistro, my previous large projected which was cancelled. The character design has been completely ripped. I also used a bunch of music files which I had downloaded for Brass Bistro that still fit the tone of Mecha Market.

Evaluation

I think my organisation of the project was a lot better than all of my previous ones because I did a lot more notetaking on what needed to be done and why I made some decisions which helped keep me on track.

Compared to Brass Bistro, I didn't really learn from my advice to not focus so much on art early on but I did avoid changing art styles too much and remaking assets unnecessarily.

I want to continue to update the game a little bit with some features that players have requested.

What's next?

For future games, I want to prototype more before starting projects so I can have more interesting or unique gameplay instead of being so derivative of other games.

I want to try music production and learning Rust probably.

I also want to revisit some older games and make them more playable or just release cancelled ones in a barebones playable state.

Play the game!

Resident Evil 4

lochi — Sat, 29 Mar 2025 24:00:00 GMT

Resident evil 4 is really good. It fails at being a horror game, more a horror themed action game. One of the biggest turnoffs for me when playing old games is how they handle death and lives, especially if it means repeating large portions of the game after failing.

Luckily it saves itself from a lives system and just resets the player to the start of the combat encounter when you die. In what was going to be re4, the original Devil May Cry, lives are handled a lot worse with you needing a special item to respawn or you have to start the whole mission over, really breaking the flow of the game.

The tank controls and pointer aiming of the Wii version add a lot of tension to each fight. The game has been tuned around the player's move set because of the control that's taken away from the player. For example, you can't move while aiming so you have to aim fast to start moving again and dodge attacks. You also can't strafe sideways so changing directions is slower.

The ammo system is also super well balanced. You can't just use a single weapon the whole game because you don't get enough ammo. You have to keep cycling through your whole arsenal and experimenting with different weapons.

There were random quick time events with motion controls that just mean waving your arms around to dodge to boulder but at least the fail animations were funny.

The writing was also so silly. Leon really says "women ☕️" at one point. The cheesy dialogue and the silly video game elements contrasted against the pretty bleak and dark environments like being presented with McDonalds toys of everyone they know speaking lines like it's been compressed through a wiimote speaker is pretty funny.

Complaints

Flash grenades are borderline useless. Only after searching up what they do did I realise they could be used against the parasites but from just playing the game I never figured this out.

Final thoughts

This was my first RE game after being a dmc fan for a while and I can totally see where the inspiration comes from. Feels like an old action movie. You should totally play it, but I would totally recommend playing on the Wii with motion controls if you can. One day I will play the remake and the vr mode (if I ever get a headset). Re4 remake vr is only on PSVR2 which is too expensive :(

Capcom also has so many good games I need to play still. Maybe okami or monster hunter? Ghost trick and Kunitsu-gami look really unique too.

Failing to finish a game

lochi — Wed, 05 Feb 2025 24:00:00 GMT

Surely in 7 months I with enough dedication I could make a decent game. Well I didn't do that.

I'm fairly new to game development, but since mid-July 2022, I've been pretty consistently building games in Godot. At the time of writing that makes it 900~ days. What do I have to show for all this work? Not that much as I would have hoped. Mostly because of the amount of projects that have been scrapped.

Brass Bistro

For 7 months, I worked on Brass Bistro, a game I look back on with some regrets.

The premise is you play as this robot who runs a restaurant, traversing alien worlds to find better ingredients.

It was hard to write a definition of the game here because my scope and ideas of what it should be kept changing throughout development. Scope creep turned what was a small game about running a restaurant into almost an exploration game.

Originally, the player would farm their own crops, then they would explore their environment to gather ingredients instead, THEN the environment was for finding permanent collectibles to upgrade the player's movement as well. I even had a hover move similar to R2D2 in the lego star wars games at one point.

For what was supposed to be about running a restaurant, I also recall spending a lot of time on building a system inspired by Crosscode that fakes 3D geometry and physics in a 2D isometric perspective. This was far beyond my skill level and took weeks and weeks to make work for something that added nothing to the core gameplay of running the restaurant. The movement then changed from simple hover platforming to just point and click walking. But hey, the fake 3D stuff looked cool.

This is the character designs progression. Surprisingly, it got lower fidelity as I tried to simplify the graphics to make development easier. Originally I even used a 3D rendered character that was projected into the 2D world.

Left : Oldest, Right : Newest

In the end, it strayed so far from the idea of a cooking game that when I went back to add the cooking mechanics, I realised my code was such a mess, it was nearly impossible to add any of the things I needed without breaking a million other systems.

The one thing I am quite happy with about this game is some of my pixel art. The main character especially is pretty cute and I plan on just copying the character design for my current project.

Looking through all my old commits though, I reckon the artstyle didn't really improve that much, but became just kinda different in each iteration.

October 2023

Wow 3D rendered character I sure hope this doesn't disappear in the next 0-3 months!

Sometime between October and January

January 2024

Short

March 2024

aesthetic pastel pallette

At at certain point, the mess was too much and I opted to abandon the project. It was structured so poorly that it was probably worth rewriting the whole thing from scratch.

What have I learnt?

So when I say I haven't made as much as I would have liked, it's because majority of the time I've spent working on games has been dedicated to projects that never saw the light of day.

If I could go back, I would have focused on participating in more game jams and doing smaller projects that were more achievable. By making games that were too big like this and a few others (see demos), I made some things that are pretty technically impressive but were too big to finish in a reasonable time.

You can see on my games page and itch.io that a lot of the actually released games were just jam games, where I had time pressure to finish the game.

Also, they aren't lying when they say to make small games. My first game was pretty well scoped. This, not so much. If I had more slowly ramped up the complexity, I reckon there would be a lot more finished games on my itch page.

I'd say, if you're taking on a big project, always be sure that you are SUPER sure that you have the skills to accomplish it, otherwise you are setting yourself up for disappointment when future you has to either spend far too long or ditch it altogether. I'm not saying you shouldn't experiment with harder stuff though, just make sure you're ready for it.

In this case, I tried to make the faux 3D, crosscode styled physics without having the technical skill to pull it off well.

Also don't focus on art early on in development. It might be fun to make the game look nice but you end up wasting time on things that might never be shown to people anyways. First make the game fun, then make it pretty.

Present day - What I'm doing different

My current project right now is my grocery store game. It's described much better on my current projects page but to summarise, its a grocery store simulation game where you run your store. I'm trying not to overscope this one too badly but this is by far my biggest project that I will (hopefully) release.

Setting a clear scope. By setting a clear scope on how large the game should be, you can avoid setting up too much work. Initially, the store you would have managed was a department store, however this means an enormous amount of items would have needed to be included to seem reasonable which increases development time without adding much to the experience. So instead, I just made it groceries which still allows for a lot of variety in items but not too much.
I'm also keeping track of my goals more efficiently by writing LOTS of to-do lists so I don't lose track of what needs to get done in order to make progress forward. If I don't track what needs to be done, I end up just fixing up really minor details over and over again so that barely anything changes like I did with the art style of Brass Bistro.

Conclusion and advice

But did those 7 months really do nothing? I learnt mistakes to avoid and improved my programming a ton. I also turned it into a blog so at least I made something out of it