last gamejam Crousteam’s (devs behind PRISM) game wasn’t judged too. I think the judges are either corrupt or just blind, because you just CAN’T skip this game by accident. @joodles03 I demand an explanation from you. It’s so dissappointing to make actually good games and just lose because judges didn’t even consider your game as an entry, and then they just don’t give you an answer and ignore you.
2 Likes
That is a MASSIVE difference in submissions since last time I joined a game jam, 500 submissions is crazy, congrats to all of the winners!
Had a ton of fun trying out the games, and creating one ourselves
2 Likes
Last year was over 1,000 so massive is an understatement
1 Like
We open sourced Cold Call Simulator, for anyone interested.
Thank you for the honorable mention.
4 Likes
@Gu4rana
Hello, and congratulations to all the winners, and as always, thank you to the Roblox Engagement Team for this opportunity.
I apologize if my message might offend anyone, but I genuinely just want to express a concern.
With all due respect to the winners and the mentioned games, it feels a bit unfair from my perspective. After looking at the selected games, I believe the game we created could absolutely compete with them. It’s very disappointing for us not to even be mentioned, especially considering that our interpretation of the theme matches many of the listed games, including one of the winners.
We worked really hard on this project and truly believed it fit the theme well. For this reason, it’s disappointing not to see it recognized among the entries.
Our game: Blackpole Protocol | Play on Roblox
Thank you for taking the time to read this message.
— Blackpole Protocol Team
6 Likes
I tried the game, It’s absolutely peak clearly deserved a ranking
4 Likes
Fire submissions, congrats everyone!!
1 Like
I don’t think they will change anything in the results. They are completely indifferent to the players’ opinions, because the exact same situation happened a year ago when we submitted the game, and no one from judges even played it. We literally didn’t notice any increase in visits, and no one answered our questions ¯_(ツ)_/¯
1 Like
yet I’d still like to hear answers about all of this from organisators of the gamejam
1 Like
@joodles03 , @Gu4rana could someone of you please answer?
1 Like
do you not think EVERYONE worked hard on their entry? they think their game deserves a mention as much as you do. i think you’re just being a sore loser
3 Likes
I honestly think my message was very respectful, I’m not saying others didn’t work hard on their entry neither I’m expecting to win or be honored after my message, I just wanted to express my concerns.
Also as I can read from other posts, some of the submissions didn’t even receive a review, so I just hope this doesn’t happen in future game jams.
I’m sorry if my message was offensive to you, but there is no reason to point the finger against anyone.
3 Likes
Félicitations a tous ! J’ai beaucoup aimé tester vos jeux et voir comme chacun a imaginé et développé tout ca , parfois surprenant , parfois quelques bugs .
J’ai aussi tenté ma chance en participant mais se n’était pas facile seul en 48h ^^ ( lundi travail)
Je redécouvre aussi Studio …
je vous la partage quand même , essayer d’arriver a la fin .
on remet ca l’année prochaine et en équipe ^^
1 Like
I never mentioned last year, I just said last time I joined.
So it is a pretty massive difference since I joined a gamejam last time, there is no need to put words in my mouth, thanks.
1 Like
I never got the avatar items 
Will take around a week to get it if eligible! 
1 Like
I really liked prism, the game had a really nice map, some nice animations it should definitely deserve atleast a honourable mention. Actually if i’m honest many games here deserved a better ranking than they actually got, like another game i liked in this gamejam called “Lumi” , the mechanic is really cool and the artstyle is also really cute but unfortunately the communication & the judge rankings in this event was just awful.
1 Like
That took ages. Anywho.
*all statements made in this post were made in jest and are entirely halfhearted and synthetic, maximally abiding by the Roblox Terms of Service and DevForum guidelines, and are intended solely to provide relevant discussion and critical analysis in the context of a Roblox game jam. this is not a personal attack on any developer, team, or Roblox employee.
I am not going to pretend that I am not salty.
Here are some of my thoughts, starting with a review of the winners and mentions. congrats.
extract first
Ethics Pending: Extract First, Ask Later. winner. Most Engaging Experience.
let’s be clear about what this game is. it’s a factory tycoon. you flatten land, place drills, route conveyors, fill storage. repeat. the loop is factorio if factorio had been lobotomized, stripped of all systemic pressure, and repackaged with a coat of vfx so thick it reads as compensation. every minor interaction gets waterboarded with sound effects and particle bursts, as if the game is aware, somewhere in its scripted bones, that the actual gameplay cannot carry the weight of your attention on its own.
the polish is real. credit where it’s due. five people worked on this and it shows in the asset quality, the building style, the general legibility of the ui. leon_kokosnuss built something that looks like a game. the problem is that looking like a game and being one are, under pressure, surprisingly distinct claims.
here is the thing about factorio, since that comparison is doing real work here. factorio is engaging because its systems are coupled. throughput pressure in one node propagates backward through the chain. you are constantly solving a dynamic, shifting constraint problem. the engagement isn’t in placing the drill. it’s in what placing the drill forces you to reconsider. ethics pending has no such coupling. the loops run parallel and non-interacting. there is no second-order consequence to any decision you make. it is a checklist dressed as a factory game, and the distinction matters enormously if the category you are being judged on is “most engaging.”
now. the theme.
“the locals are rocks.”
that’s their joke. that’s the entire thematic engagement with “first contact.” a pun. an ironic shrug. the description reads: “this is not exploitation. it’s efficient cooperation.” which is funny, maybe, in the way that a three-second joke is funny, but it is absolutely not a game about first contact. the theme is set dressing. it is the premise of the title and then it evaporates. the rubric explicitly weights “usage of the theme” equally against every other criterion. equally. so either the judges decided that ironic deflection counts as thematic engagement, or they didn’t actually apply that weight the way the terms said they would. both possibilities are uncomfortable to sit with.
but here is the part that made me actually put my head in my hands.
THE ENTIRE GAME IS CHATGPT TEXT. the title. the description. the tutorial dialogue. the companion robot’s lines. all of it. every single word of in-game text is shamelessly, obviously, generatively produced. no voice. no perspective. no taste. it reads like someone typed “write funny space mining tutorial dialogue” into a chatbot and hit publish. which, to be fair to them, is probably exactly what happened.
the rubric asks for “fun, surprising, immersive.” surprise, specifically, requires a perspective. it requires a writer who has chosen something unexpected. ai-generated dialogue is, by construction, the average of all prior dialogue. it cannot surprise you because it is assembled from what you already expect. every line in ethics pending has been pre-digested into familiarity before it reaches you. the surprise budget is zero before the game even starts.
and the ux. the unskippable dialogue. the slow startup drowning in visual noise. “keep players wanting to come back for more” is the exact criterion, verbatim from the terms. it is very hard to want to come back when the front door has a mandatory loading monologue and the first three minutes are tutorial friction.
the honest structural read of why this won is not complicated. a fatigued judge with limited time sees: conveyor belts, polish, tycoon loop, vfx on every interaction, five-person team, clean thumbnail. the rubric checkbox fires. genre legibility did the scoring. the actual experience of the game, its thematic abdication, its ai-slop text, its shallow loop, its ux friction, requires sustained time to notice. judges, grinding through 500 entries, did not have sustained time.
that’s not a complaint about the judges as people. it’s a complaint about what the incentive structure of jam judging selects for. legibility. the game that looks most like the platonic ideal of its genre from the outside. not the game that actually does something.
ethics pending exists to be a game. it achieves that. it is a game. and in a tournament that rewards looking like a game from a distance, that turns out to be enough. congrats.
dye
dye. Most Creative Interpretation of the Theme. let’s actually look at what that criterion says.
“boldest and most imaginative understanding of the theme.” “transform the theme into something unexpected and memorable.” those are the words Roblox put in the document. bold. transform. not “make a nice game.” not “have good color theory.” transform the theme into something unexpected.
so. what is dye’s interpretation of First Contact?
you are a colorless character. you touch color. color lets you through gates. you mix colors. you give color to a pile and it becomes a mushroom. at the end, you deliver color to a static T-posing figure and it animates.
that’s it. that is the full semantic relationship between this game and the concept of First Contact. “contact” as in… you touch things. and they change. which is, if we are being extremely charitable, a reading of “contact” as property exchange. you make contact with a red fire, you become red, the red lets you through a red gate. first contact as a mechanic of permeation.
i want to be clear about something. that is not bold. that is not a transformation of the theme. that is a color puzzle game that noticed the word “contact” exists inside “First Contact” and called it a day. the “first” does zero work. there is no significance to the first contact event. no weight. no setup. you touch the first dye at the start of the game and nothing about it reads as momentous or thematically loaded. it’s a tutorial. it teaches you the mechanic. the theme is not being interpreted, it’s being decorated around.
the ending is the one moment where something almost happens. you deliver color to a static T-posing character and it animates. which is, genuinely, a cool image. contact as the gift of motion. contact as what separates the inert from the living. if the entire game had been built toward that image, consciously, with that thematic payload in mind, dye might have had a case. but it wasn’t. the T-pose figure appears at the end of a puzzle sequence that was about gates and mushrooms and fire. the resonance is accidental. a stopped clock.
and then there’s the “bold” question, which is where this gets actually frustrating.
boldness implies risk. it implies a design decision that could have failed, that required the developer to commit to something uncomfortable or strange or alienating. dye is the structural opposite of bold. it is cozy. intentionally, aesthetically, deliberately flat and inoffensive. the color palette is gentle. the puzzles are calm. the pacing is unhurried. there is not a single sharp edge anywhere in the experience, and i mean that both formally and thematically. this is le cozy millennial-developed game. the kind of thing that gets passed around with “this is so charming” energy on indie dev twitter and then forgotten in six weeks.
which, again, fine. there is nothing wrong with making a calm, cohesive, well-executed puzzle game. kenami and Soorbeet clearly have taste and the mechanical coherence of the color system is real. the downstream consequences of dye states, mixing colors to produce effects, giving color back to environmental objects, that’s a properly designed system and it shows.
but that is a “Best Technical Quality” argument. maybe “Most Innovative Gameplay” at a stretch. it is not a “Most Creative Interpretation of Theme” argument. those are different axes and this game was placed on the wrong one.
what actually happened here is legible if you zoom out. 500 entries. fatigued judges. dye is immediately, visually, unmistakably cohesive. it looks finished. it feels considered. the aesthetic is consistent from the thumbnail to the last puzzle. a judge with limited time sees this and the pattern-match fires: “this feels creative, it has a distinct visual identity, the color thing is clever.” and it is clever, as a mechanic. but clever mechanics are not thematic interpretation. the rubric asked for the game that most boldly transformed the concept of First Contact into something unexpected. out of 500 submissions. and this is what landed.
…okay. 500 games. maybe every single other entry was somehow worse. maybe the field was genuinely that thin on thematic ambition. but dye’s presence in this specific category, above whatever else was in that pool, suggests the judges were grading on overall quality and aesthetic coherence and then retroactively assigning it to whichever category it fit least badly. which is a failure mode of jam judging that deserves to be named plainly.
stellar supervision
Stellar Supervision. Best Technical Quality.
the criterion, again, verbatim: “optimization, quality of mechanics, absence of bugs, smooth, reliable gameplay that demonstrates technical mastery.”
technical mastery.
let’s talk about what mastery actually means, because apparently this needs to be said. mastery is not “the game runs.” mastery is not “the tweens are smooth.” mastery is not “the sound triggers on the correct frame.” those are baseline competencies. those are the minimum viable bar for submitting something that doesn’t embarrass you. half of those 500 entries probably hit that bar. the question the category is asking, or should be asking, is: which game demonstrates command of the engine at a level that is genuinely difficult to replicate? which game has systems under the hood that required real engineering decisions?
Stellar Supervision’s answer to that question is: connect four with an NPC, a simon-says color memorization minigame, and some environmental interactions that trigger when you do the thing.
i want to be fair. the game looks incredible. doomforks and Justine_Emmerson and reamnos built something visually coherent and genuinely impressive for three days. the asset quality is real. the animations are fluid. SComics did real work on the sound design. the modellers and artists on this team are legitimately talented and the result shows.
but here’s the thing. and this is the crux that the judges apparently decided not to engage with.
NONE OF THAT IS TECHNICAL QUALITY IN THE SENSE THE RUBRIC DEFINES IT. a billion drawcalls and a beautiful rig do not constitute technical mastery of game engineering. you could fire the scripter, hand the codebase to a sufficiently prompted language model, and the gameplay would be functionally identical. connect four is a solved game with a solved implementation. simon says color memorization is a solved game with a solved implementation. the “haha funny environment interaction” moments are event triggers. this is not difficult scripting. this is not systems architecture. this is not anything that required PuffoThePufferfish to make a single non-obvious engineering decision.
what the judges rewarded here is production value. which is a real thing and worth rewarding, somewhere. but production value and technical quality are not the same axis and conflating them is a categorical error that the rubric explicitly does not support. “optimization, quality of mechanics, absence of bugs” is an engineering checklist. Stellar Supervision passes it the same way 250 other entries probably passed it, by not breaking and looking nice while doing so.
and the theme. oh, the theme.
“You are the first human to make contact with Aliens; unfortunately, they have a job interview to attend.”
that’s the joke. that’s the full thematic engagement. first contact as a comedy bit. you are told, in the opening exposition dump, that first contact has happened. it happened before you loaded in. the game begins post-theme. you are then handed a connect four board. the “first contact” is not something you experience, it is something that is narrated at you so the game can move on to its actual content, which has nothing to do with it. the spoonfed exposition just happened to contain traces of “First Contact” and that was enough.
this is the second winner in a row where the theme is a punchline rather than a premise. “the locals are rocks.” “they have a job interview to attend.” both games treat First Contact as a setup for a wink at the camera and then immediately evacuate the thematic territory. and both placed. which tells you something about how seriously the theme criterion was actually weighted against the overall impression of the entry.
Stellar Supervision is a beautiful game. it is a technically adequate game. it is not a technically masterful game in any sense that the category was supposed to reward. what it is, plainly, is a game with a great modelling team and a coherent art direction and enough visual polish that a fatigued judge pattern-matches to “most technically impressive thing i’ve seen today” without stopping to ask whether the mechanics under the surface required any meaningful engineering at all.
the judges saw the billion drawcalls. they said yes. nobody checked again.
on air
On AIR! Best Use of Interactive AI.
the criterion: “show us how AI can amplify your vision through new ways for players to interact with the world, communicate, or manipulate the environment. remember it isn’t about using the most AI features, but how to use them in ways that are meaningful and impactful.”
meaningful and impactful.
let’s establish what On AIR! actually does with AI. the game uses speech-to-text. you read lines into a microphone. the game scores you on whether you said the thing. that’s it. that is the full extent of the AI integration. there is no generation. there is no adaptation. there is no system that responds to what you said in any way that required intelligence to implement. the speech-to-text is a input modality swap. it replaces typing with speaking. the game underneath is identical whether you typed it or said it, except typing would be more precise, more accessible, more reliable across hardware configurations, and frankly more fun because it has more surface area and moving parts than praying your microphone catches the syllables correctly before the timer ends.
the argument being made implicitly by the judges is that using speech-to-text as the input method constitutes meaningful and impactful use of AI. which, if you accept that framing, then every game that puts a microphone check anywhere in its loop qualifies. the bar is not “AI amplifies the vision.” the bar becomes “AI is present in the game somewhere.”
and then there’s the theme.
the comic is genuinely charming. the art is good. Sheenyo and Skarletbun clearly have real illustrative talent and the visual identity of glebs-TV has personality. the joke, “first contact” as “first one we called,” is actually a creative reframe in isolation. it’s the kind of wordplay that should lead somewhere interesting.
except it doesn’t lead anywhere. the joke lands in the comic panel. you read it. you appreciate it. and then the game begins and the theme is gone. you are on a stage performing microgames. the aliens are the audience. the premise evaporates into set dressing the moment gameplay starts. the “first contact” is, again, an exposition delivery mechanism and then the game moves on without it.
this is now the third winner in a row where the theme exists entirely in the framing text and not in the experience of playing the game. the comic tells you the premise. the description tells you the premise. the game itself is a microgame performer that would be functionally identical if the aliens were replaced with humans and the setting was a talent show on earth. the theme does not shape the mechanics. it decorates them.
but the AI thing is where i actually lose my mind a little.
Roblox gave developers five new AI tools for this jam. speech-to-text, text-to-speech, text generation, 3D generation, 4D generation. five tools. the category criterion says “it isn’t about using the most AI features, but how to use them in ways that are meaningful and impactful.” which is an interesting thing to write in your rulebook when the winner used exactly one feature and used it as a direct input replacement for a mechanic that would work better without it.
meaningful and impactful. the speech-to-text in On AIR! is meaningful in the same way that a microphone is meaningful in karaoke. you need the microphone for the karaoke to work. that does not make the microphone a creative use of audio technology. it makes it a hardware requirement.
compare this to what a meaningful use of these tools could look like. text generation that responds dynamically to player decisions and alters the narrative state. speech-to-text that captures what you say and not just whether you said something, feeding that into a system with real downstream consequences. 4D generation that produces environmental objects that react to player input in ways that weren’t pre-authored. those are uses of AI that would be difficult or impossible to replicate by swapping the feature out. On AIR!'s speech-to-text is trivially replaceable. the game explicitly admits this in its own description: “The AI Interaction in-game is only speech-to-text, no other forms of AI are used.”
they wrote that sentence as a transparency note. it reads, in context, as an accidental confession.
i didn’t play far enough to be exhaustive here and i’ll own that. but what i saw was a game that found the thinnest possible scaffolding onto the AI category, dressed it in good art, and submitted. and it won. which means either the other entries in this category were genuinely worse at AI integration than “speak into microphone,” or the judges were, again, pattern-matching on overall impression rather than criterion-specific evaluation.
both possibilities say something worth saying.
first rift
First Rift. Most Innovative Gameplay.
“genuinely new and groundbreaking.” “complex, multi-layered progression systems.” “blending genres in unexpected ways.”
let’s establish what First Rift actually is. it’s a top-down shooter. WASD to move, mouse to aim, click to shoot. enemies spawn from rifts and slide toward you. you shoot them. you walk to the next barrier. repeat. at some point you press Q to switch between two characters, one ranged, one melee. that’s the mechanic. that’s the full extent of the innovation on offer here.
Q to switch character.
i need to sit with that for a second because the category criterion uses the word groundbreaking. groundbreaking. as in, this breaks ground. as in, ground that was previously unbroken is now broken because of what this game does. the dual-character swap mechanic has existed in games since roughly the paleolithic era of game design. it is in Gunstar Heroes. it is in every action RPG with a class system. it is in games that were old when the developers of First Rift were born. this is not a new mechanic. this is not a blend of genres. this is a top-down shooter with a swap button.
and the enemies. the enemies are the part where i start to genuinely question what the two scripters were doing for 72 hours, because the enemy AI is a WalkTo in a loop. that’s it. that is the complete implementation. they detect you, they move toward you, when close they deal damage and play a delayed animation. there is no attack pattern. there is no variation. there is no systemic behavior that creates interesting player decisions. the “gameplay” of First Rift is: run in a circle. shoot the thing that chases you. the overheat and ammo mechanics are present but so unexposed and poorly communicated that they register as noise rather than depth. you discover them by accident and they don’t change how you play in any meaningful way.
five people made this. five. MarkMuz123 on building and lore. lethrjakt on building and texturing. Xelixis on scripting and GUI and VFX. catmanguy_s on scripting. fineeena on art. two scripters. two full scripters on a game whose enemy behavior is a single pathfinding call in a loop and whose primary mechanic is a character swap. i genuinely do not know what the second scripter’s git history looks like and i am not sure i want to.
the theme is where the last remaining structural support collapses entirely. First Contact in this game means: you walk up to an NPC. dialogue plays. he asks for your help. you fight him. you can then play as him. that’s it. the “first contact” is a five-second cutscene framing device before the game reveals that it was a top-down shooter the whole time. the theme doesn’t shape the mechanics. the theme doesn’t shape the progression. the theme is a napkin the actual game was sketched on and then discarded.
the story, to its credit, gestures at something. there’s potential in the setup. and then it is never developed again. ever. the lore exists as texture, not as substance. you get the premise and then you get wave enemies from rifts for fifteen minutes until you stop playing.
the game has combat. combat is a legible form of gameplay loop a fatigued judge can evaluate quickly. you see enemies, you see shooting, you see health bars and VFX on hits, and the pattern-match fires: “this is a game with systems.” the Q-swap mechanic registers as a distinguishing feature in the first three minutes because it’s the one thing that isn’t identical to every other top-down shooter. and three minutes is roughly the evaluation window a judge has per entry when they’re moving through 500 of them.
“genuinely new and groundbreaking.” that phrase is in the official terms. roblox wrote that down and then awarded the category to a top-down shooter where the enemies walk in a straight line toward you.
i don’t know what else to say about that. the terms said what they said. dude.
honorable mentions:
love in the air
i honestly have no idea what is going on here.
Love in the Air. honorable mention.
the pigeon shoots things near a plane. that’s the game. you move a pigeon, you click, obstacles die. the girlfriend got kidnapped. the plane needs protecting. the loop has no second layer, no emergent decision, no moment where you discover something the tutorial didn’t tell you. it is the complete and total product. a pigeon. shooting. things.
the thematic connection to “First Contact” is that the pigeon falls in love at first sight. that’s it. they looked at the phrase, extracted the word “first,” pointed it at a romantic subplot, and submitted. this is not a creative interpretation. this is a ctrl+f on the prompt document.
the honorable mention list is not a participation ribbon. it is not a “good effort” sticker. it is not the judges gesturing warmly at a beginner team and saying “we see you, keep trying.” it is supposed to identify the games that nearly won a category. the near-misses. the entries that had a genuine, articulable case for placing and lost on a narrow margin. that is the structural function of an honorable mention in a competitive judging context. it means: this was close. this almost got there.
Love in the Air was not close. it does not have a case for any category. not engaging, the loop has no depth or consequence. not creative, the theme is a ctrl+f on the word “first” inside a love story that has nothing to do with First Contact as a concept. not technical, there is no engineering here that required a non-obvious decision. not AI, there is no AI. not innovative, this is a genre that has existed for forty years and this entry does nothing to it.
the team being beginners is not a mitigating factor in this specific context because this specific context is not about beginners. the honorable mentions are not a junior division. there is no category for “most complete game from a team that was learning as they went.” if that category existed, Love in the Air might belong there. it does not exist. what exists is a competitive rubric with named criteria, and this game satisfies none of them at the level that “nearly won” implies.
finishing a jam game as a beginner is fine. it’s a real thing. go be proud of it in the appropriate thread. but the honorable mention designation is a judge saying, on the record, that this game belongs in the top eleven of five hundred submissions by the standards the rubric defines. and it doesn’t. it isn’t close to belonging there. its presence in that list is not a close call that went the wrong way, it is a selection that reveals the rubric was not the actual decision-making instrument. something else was. and whatever that something else was, it produced this.
a pigeon. shooting. things. honorable mention.
the errant machine
The Errant Machine. honorable mention.
CheezBoy27 made this alone. that’s the first thing worth saying. one person, 72 hours, underwater, and the result is something i would describe as tasteful, which is a word i did not use once across the entire winner list and mean it without qualification.
the mechanic is rotating part puzzles where some components only articulate in one direction. it’s flat. it’s consistent. it never becomes annoying, which is a design achievement that sounds trivial until you’ve played puzzle games where the central mechanic outstays its welcome by minute four. it doesn’t do that. it knows what it is and stays there.
and the world is beautiful. genuinely beautiful in the way that matters, not production-value beautiful, not asset-budget beautiful, but considered. the underwater scenery has lore implications that live in the geometry and the lighting and the ambient sound rather than in a dialogue box. you feel the weight of something larger than the immediate puzzle. it has atmosphere the way that only happens when someone made decisions about what not to include.
so why is it here and not higher.
because it didn’t optimize for anything. no hook that fires in thirty seconds for a fatigued judge. no category it dominates. the theme residue is minimal, present as texture rather than premise. the gameplay loop has no second layer, no escalation, no moment where the system reveals it was deeper than it looked. it is a good game that does not behave like a competitive jam entry, which is the same failure mode i will be describing shortly in my own post-mortem, so i say this without any superiority. it is a game built to be experienced rather than evaluated. those are different products and only one of them wins rubric-based tournaments.
good job, ChezeeBoy27. genuinely. go read the guidelines next time. and then possibly ignore them and make this anyway, because this is better than most of what won.
busy line
Busy Signal. honorable mention.
you are a phone operator. you plug cables into sockets. incoming call wants department B, you find the B socket, you plug it in. that’s the game. that is the complete and total product.
the old telephone switchboard aesthetic is committed to, i’ll give it that. the visual language is consistent. someone at Underlogic Studios has a hyperfixation and it shows in the specificity of the prop design.
but there is no game here. there is a simulator of a task that was automated out of existence in the mid-twentieth century, which would be interesting if the simulation went somewhere, if the calls escalated, if the mystery of the “mysterious company” the description gestures at ever materialized into something you actually experienced rather than something hinted at in two words of flavor text. it doesn’t. you plug cables. the calls resolve. you plug more cables.
First Contact is not here. i looked. “first contact” as in “the first time you contact a department by cable”? maybe. if you extend enough charity to strain something. the honorable mention rubric implies this was nearly a category winner. i cannot construct the argument for which category. not engaging, the loop has no consequence or escalation. not creative, the theme is either absent or so laterally interpreted it disappeared. not technical, not AI, not innovative. stale crackers. fine if you’re hungry enough. nobody should have been this hungry.
i don’t know why it’s here. moving on.
the rest i either did not play, do not remember and do not want to play, or have nothing significant to be said about them. if you want my review or elaboration on something, just ask
^ moreso directed at the judges. the games were competent, but apparently neither the gamejammers nor evaluators read the guidelines. 500 games is a lot. i am sure there were better candidates.
thoughts^2
not a single winner meaningfully engages with the theme in the experience of playing the game. every single one of them either evacuates “First Contact” into an opening comic panel, a description blurb, a piece of tutorial dialogue, or a winking pun, and then immediately moves on to being a game that has nothing to do with it. Stellar Supervision is the only one where the thematic premise is at least legible in the aesthetic. aliens. first contact. yes. fine. you can draw a straight line between the concept and the setting without needing a decoder ring. except the actual first contact event happened before the game loads. you are told about it. and then you play connect four.
so across five categories, five winners, the theme “First Contact” was treated as: a pun about rocks (Ethics Pending), a color puzzle mechanic with “contact” extracted from the phrase (dye), a pre-game cutscene (Stellar Supervision), a comic panel about a phone call (On AIR!), and an NPC dialogue trigger (First Rift). the “usage of the theme” criterion was equally weighted against every other scoring axis, verbatim in the terms and conditions, and the collective answer from the judging panel was: yeah, this counts.
and then there’s the category satisfaction question, which is the part that actually makes me want to put my head through a wall. out of 500 submissions. five hundred. these were the games that most exemplified Most Engaging, Most Creative, Best Technical, Best AI, Most Innovative. a factory tycoon with AI slop text and unskippable dialogue. a color puzzle game that accidentally gestures at its theme in the final cutscene. a visually stunning tech demo with connect four underneath. a speech-to-text wrapper around a microgame. a top-down shooter where the enemies walk in a straight line.
i have played games from previous roblox jams and genuinely admired them. genuinely. not in a “good for a jam” way, in a “this is a real game with real ideas” way. i did not feel that once across this entire winner list. i felt competence, sometimes. i felt production value, occasionally. i felt the ghost of an interesting idea that was never developed, repeatedly. i did not feel anything that made me think: yes. out of five hundred games, this one. this is the one that deserved it.
maybe the field was thin. maybe the 500 submissions were, on aggregate, genuinely weak and these were the best available options in a bad pool. i don’t have all 500 cached. i cannot say with certainty. but what i can say is that five winners, reviewed honestly, each fails to satisfy its own category criterion in at least one significant and non-trivial way. and if the judging was rigorous, that shouldn’t be possible. criteria exist precisely to prevent vibes from doing all the scoring. the criteria did not prevent this.
my reasoning and judgement here absolutely exist outside of simply losing and being salty. that much i am sure about. salty and correct are not mutually exclusive. (not to frame that i am this salty—it really sounds like that but really, after 20+ days (compared to the regular 15), the adrenaline has worn and i entirely anticipated my own outcome—but not this particular selection.)
my experiences
read
i went into this with a mental model so clean it was basically insulting to the difficulty of the problem. the rubric was a known constraint set. the categories were five labeled boxes and i had a working theory for how to fill each one. i am a cracked scripter, i mean that without any particular humility, and the plan was simple: acquire people to cover the dimensions i couldn’t parallelize, build something with actual architectural intent, and collect the money.
the failure was not technical. let me be explicit about that because it is the part that matters most and the part most likely to get lost in the rest of this. there was no architectural error. no conceptual mistake. no moment where i looked at the code and thought “this is wrong.” the failure was people. purely, entirely, humiliatingly, people.
i am a human being, which means i required sleep, which means the 72-hour clock was never actually 72 hours, it was 72 hours minus whatever the biological minimum was to keep my hands on the keyboard without the output becoming incoherent. that constraint is fixed. cannot optimize around it.
and then nobody showed up. the bottleneck was people. half of the premise was to parallelize the work i do not specialize in so that i don’t choke on serial amateur tasks.
i want to be precise about this because “the modellers flaked” makes it sound like a narrow failure with a narrow fix. i reached out across multiple roles to a lot of people. not a few. a lot. some ghosted immediately. some expressed interest with the specific energy of someone who has no intention of following through and then went quiet within hours. some i think simply did not take the pitch seriously, which is understandable in the abstract and genuinely infuriating in practice. i had platform presence. i had, technically speaking, connections. i sure as hell had connections. what i did not have was friends, as in people who would pick up the phone when it mattered. the communities i tried were worse than useless. cold outreach for an unpaid 72-hour jam is a structurally weak ask and i knew that going in and tried anyway because the alternative was going in alone, which is what happened regardless.
so the plan, which had been “assembled team builds a bludgeoning good game with genuine passion and then extracts the prize,” became “one person does all of it.” not as a contingency. as the only available path forward. one brain. two hands. a caffeine budget. and 72 hours that were already not 72 hours.
before i go further: i should say what the original ambition actually was, because the gap between that and what shipped is the part that still stings more than the loss itself.
the rubric had five categories. the strategy, before it collapsed, was to target all of them simultaneously with a single coherent concept, not by spreading thin across five unrelated ideas but by building something whose individual systems each spoke to a different criterion. most engaging: a tactile navigation and extraction loop, hull damage, physics-enabled ring debris, a vulnerable “reach window” mechanic that made measurement feel expensive. most creative interpretation: first contact not as a greeting but as a causal event, the player’s first mundane instrument calibration being the contact, the anomaly having already noticed them before the dramatic midgame reveal. best technical quality: a diegetic cockpit with no floating UI, custom camera math, spatialized audio, a phase controller that prevented sequence breaking, zero friction in the first sixty seconds. best use of AI: ShipSystem as a narrative hijack vector, TTS voice that maintains identical cadence while the intent underneath shifts, STT for hands-free operational commands, text generation restricted to structured protocol cards rather than open chat, the whole thing engineered specifically to avoid roblox’s 18+ age gate without neutering the mechanic. most innovative gameplay: making observation costly, every time you looked at the anomaly it looked back, the subsequent navigation degrading in specific ways that felt like consequence rather than punishment.
that was the blueprint. it was coherent. it mapped onto the rubric’s own language with more fidelity than i think any of the winners managed.
none of it shipped intact.
what i actually delivered was, by my own standards, utterly terrible. i say that without performance. the game ran. it held together. it had a spine and a thesis and a reason to exist that extended beyond “it is a game that contains game elements.” but the granular execution was a wreck. features fell away one by one as the hours collapsed, not because the vision was wrong but because one person serialized through every task that a team runs in parallel hits a hard ceiling eventually, and i hit it repeatedly. no asteroid navigation. no survival loop. no hull repair. the gameplay that was supposed to justify “most engaging” got reduced to a story game where you click here and here and pay attention. i watched it happen in real time and could not stop it because stopping meant not finishing the other half.
the audio is its own small catastrophe. i had slaved over the sound design. forced myself to use the new audio API specifically to satisfy the technical category. it fried everything. the sound designer i had reached out to never followed up. i kept the audio, decided i would fix it, and did not fix it, because there was no time, so the game shipped with every sound asset screwed in ways i was fully aware of and submitted anyway because silence was much worse.
the AI integration, which was supposed to be the category i dominated hardest, ended up as the most painful gap between intention and execution. roblox’s tooling was uncooperative in ways i could not resolve within the remaining time after everything else had eaten the schedule. what shipped was a bolted-on approximation. functional. not what it was supposed to be.
i submitted the final patch with seconds remaining. the last thing i touched was the parallax timing math on saturn, adjusting values by hand with no certainty i hadn’t broken something downstream. i hadn’t. the game held. the adrenaline cut out the moment the submission confirmed, replaced immediately by the specific chest tightness and light-headedness i had been ignoring for two days. basic anxiety. sounds more dramatic written down than it felt at the time.
the game satisfied the latent spine of the vision. the thesis was intact. the idea of STARSPLITTER, first contact as an epistemic boundary event, the player as an observer whose observation is itself the contact, the anomaly as something that was never waiting to say hello but waiting for a thinking system to look at it, that survived. what did not survive was almost everything granular. and i knew this. i knew exactly how far short of my own bar it landed.
and i still thought it might place.
i want to be honest about that because it is the shameful part. after twenty days of waiting, after telling myself i had fully priced in the loss, after intellectually understanding every structural reason why atmospheric narrative games underperform in rubric-based judging, i still thought it might place. i still believe i was nearly right. i ran the numbers on previous jams and predicted a general drop in developer competence for 2026. a peer reviewer called it immersive, called it one of the top spaceship games in the jam. in the context of their other reviews, that read as signal.
then the list dropped and it wasn’t there. not in the winners. not in the honorable mentions.
the part that does not resolve into something comfortable, no matter how much charity i extend toward the process, is this: i built something terrible by my own standards. half-finished, broken audio, gutted gameplay, a fraction of the original vision. the failure was not a mystery and it was not bad luck. it was the hard throughput ceiling of one sleep-deprived human with no team, doing every job in sequence that five people were supposed to do in parallel. that is the whole story. people. just people.
and the games that beat it, across five categories, with five full teams, with seventy-two hours of parallelized development, still somehow failed to satisfy their own rubric criteria in ways that are specific and nameable and not a matter of taste.
salty and correct are not mutually exclusive. i am both. i was also, genuinely, not good enough. those two things coexist without canceling each other out, and i think that is the most honest thing i can say about this entire experience.
i’m going back to working for PNEUMA on Apocrypha. which lost 72 potential developer hours to this jam. and the flowstate it took with it.
other remarks
i agree completely
i must say that i completely agree. i played through Blackhole Protocol and it was phenomenal, and its absence from both the winners and the honorable mentions is genuinely baffling.
the core mechanic is immediately legible and immediately interesting, which is a combination the actual winners largely failed to achieve simultaneously. a handheld device, left click for south pole, right click for north. magnetic attraction and repulsion applied to environmental objects and to yourself. swing into geometry, shatter glass with momentum, solve spatial puzzles by understanding pole relationships rather than memorizing a sequence. it has rough edges. it shows its jam origins in places. but it is good. nearly tasteful, actually, which is a bar i did not clear once across the entire winner list. i actually enjoyed this game. it felt like a game jam winner. it should have been one.
the technical execution deserves a specific mention too. the facility ambience runs through what sounds like roblox’s TTS pipeline. it is a small thing, non-pivotal, easy to miss. but it is AI used as atmosphere, dissolved into the world rather than slapped onto the front door as a feature. it is, unironically, a more considered use of the tooling than what won the AI category.
and the mechanic is genuinely fresh. magnetic pole logic applied to a first-person puzzle space is not a genre i have seen before in this engine. the depth compounds as you understand it. the decisions it creates are real decisions. that is what “most innovative gameplay” is supposed to be selecting for and it is sitting in the submissions pile unmentioned while something else collected the trophy. even some of the other awards.
i cannot make that make sense.
overall
mid gamejam :(
five categories. five labeled boxes with explicit, written, equally-weighted criteria. a theme that was supposed to be load-bearing. 500 submissions. twenty days of evaluation. and at the end of it, a winner list where every single entry evacuated the theme into a description blurb or a comic panel or a winking pun, and then spent the rest of its runtime being a game that had nothing to do with First Contact. a rubric that said “technical mastery” and rewarded production value. a rubric that said “groundbreaking” and handed the trophy to a swap button. a rubric that said “meaningful and impactful AI” and gave it to a microphone check that the developers themselves felt compelled to apologize for in their own description.
the AI category is the one that sits the most uncomfortably. Roblox built these tools. they shoved them into the jam categories. they wrote a criterion that explicitly said it was not about using the most features but about using them meaningfully. and the winner used one feature as a direct input swap for a mechanic that would function better without it. i would wager that the devs behind On AIR! would never dare touch an LLM outside of a jam context, out of some combination of purism or virtue signaling, which makes their win in this specific category a particularly exquisite irony. the teams that actually pushed the tooling, that tried to make the AI do something the criterion was gesturing at, built something too architecturally complex to survive a three-minute evaluation window. the teams that won did not win because they used AI well. they won because they used it safely, visibly, and without friction. the rubric said meaningful and impactful. the result said present and inoffensive.
mid gamejam. a bad one. perhaps a corrupt one. mid. competent selections from a process that was optimizing for something adjacent to but not identical to its own stated criteria. the rubric existed. the judges had it. it did not end up being the actual decision-making instrument, and the delta between what it said and what it selected is visible, specific, and not a matter of interpretation.
i had a better game in my head than i had on the screen. the screen version lost on its own merits and i own that. but this particular set of winners, against this particular rubric, in this particular jam, deserved exactly this review. do not take offence.
all statements made in this post were made in jest and are entirely halfhearted and synthetic, maximally abiding by the Roblox Terms of Service and DevForum guidelines, and are intended solely to provide relevant discussion and critical analysis in the context of a Roblox game jam. this is not a personal attack on any developer, team, or Roblox employee.
if you wish to debate, do omit my alleged hostility and be so generous as to not reduce my entire reasoning to “mad cuz lost”
congratulations. thank you for coming to my ted talk
8 Likes
Well since you wrote all of this I’m surprised you didn’t mention UNTOUCHABLE (my game). But you did say I could ask so…
1 Like
Thank you!
And I will say if I am absolutely guilty of making puzzle games with “a central mechanic that overstays it’s welcome” previously. I had intended to make a lot more puzzles in the game originally, and I knew the pipes were relatively simple puzzles which is why I only added a few. It’s great to hear that it doesn’t get annoying to the player, I’m glad I went with as many as I did. Initially I was very concerned the rotation mechanic wouldn’t be “challenging enough,” which was a dumb concern to have in hindsight because I have made harder puzzles before that people give up on because the mechanic gets boring or they assume the game broke. (Major downside to being a small dev on Roblox: if something looks like it won’t work, people generally think it’s broken because they assume there is no quality assurance on games with a handful of likes…)
Genuinely, I’m so flattered. One of my biggest peeves in Roblox games is when a game bombards you with a 5 minute unskippable dialogue lore dump before I even know if I want to care about the game. I was extra careful to make sure I didn’t go too crazy in the beginning with textboxes (besides the skippable first menu), so that hopefully people who make it to the end are motivated enough by the world to want to care about the last few dialogue boxes.
And this is completely fair, I didn’t have time to introduce new mechanics or do anything to make the gameplay (is it even gameplay? you just swim around…) interesting besides just whatever puzzle you encounter on the way to the next one. It’s definitely not a conventional game jam entry in the slightest, there’s no loop or real gameplay besides walking through a linear map.
For last year’s Eurojam I had made a game that was a more conventional, it had rounds, a weapon, a map with lots of space for combat, but it ended up feeling pretty generic. Ultimately I would love to make a game like Running From The Internet or Project Remix, but I know I don’t really have anything to bring to the table in that style of game, so I think if I do go into next years jam I am going to continue to do this style of atmospheric linear “story” game (even if it isn’t engaging to the judges). I think the majority of games like this on Roblox can be cringey, simplistic, or lacking in atmosphere, and I would love to beat this trope someday.
I am absolutely keeping these ideas in mind, all of my games would seriously benefit from a second hook or escalation.
Thank you for the genuine feedback! 
2 Likes