Show HN: Ghost Pro Racing 3D Async Multiplayer Arcade Racing in HTML5

https://gamesnacks.com/games/52v00umba6lko

3•andvari_bekho•4mo ago

Hi HN!

I built a 3D arcade racing game in HTML5 using PlayCanvas.

It features arcade-style physics inspired by OutRun, with PS1/Dreamcast-style graphics. The multiplayer is asynchronous: you race against the “ghosts” (replays) of up to 7 other players, and after finishing a run you can share a link with a friend so they can race your ghost 1v1. The game currently has 20 tracks, plus a “daily challenge” mode with procedurally generated tracks.

We’ve tested it on both desktop and mobile. It supports touch and keyboard controls, though I personally prefer playing with a keyboard in landscape mode. I’d love feedback on the controls, performance, and overall fun factor. And I hope to see your names on the leaderboard!

Comments

bigjobby•4mo ago

This is ridiculous! Love it. Already shared on my socials

andvari_bekho•4mo ago

Thanks for sharing! I really appreciate it, glad you’re enjoying it :)

gus_massa•4mo ago

Nice, it defectively reminds me of OutRun.

4:09.88

Add a few fake slow cars, because otherwise it's too lonely :(

It's nice to try and later build the profile, but I think it should be easier to discover.

gus_massa•4mo ago

Too late to edit: More feature requests:

From some TrackMania videos, you can play against with your own last/beast ghost.

I guess you are picking the 7 fastest car or the last 7 cars (that are mostly hardcore users), so they are quite fast. Is it possible to add a few slow cars, perhaps select a mix of times?

Bonus imaginary points for:

* my car obliviously :)

* the fastest/last 7 cars

* 6 cars selected with an even distribution. If the fastest is 1 minute and the slowest 4 minutes, let's say approximately 1:15, 1:45, 2:15, 2:45, 3:15, 3:45

* 2 cars with a time close to the expected time of the player. If in the last round the player got 2:17, let's pick an old car with 2:07 and another with 2:27, so the player has two ghost nearby.

andvari_bekho•4mo ago

Hey gus_massa, thanks a lot for the feedback! Those are some really cool ideas.

For the other players you see in the race, we’re currently picking random players from the database. It would definitely be nice to select a mix of fast, average, and slower cars for a more balanced distribution (that was our original plan), but unfortunately the GameSnacks API doesn’t support that feature yet. Since some hardcore users play a lot, their ghosts tend to show up more often in the pool.

Right now you can already see your own ghost car in the Daily Challenges, but being able to see your “best time ghost” in the standard races is also a really cool idea!

gus_massa•4mo ago

I guess ypu are picking 7 random users. Is it possible to pick 1000 and then automatically cherrypick 7 of them? Does the API save all the races or only the best of each player?

andvari_bekho•4mo ago

The API saves all races, now just your best one

Yes, selecting 1000 and then cherry-picking would be a great idea in theory, but ghost data is quite heavy (around 300–400 KB per ghost). We are already compressing it as much as possible: saving only about 10 frames per second and interpolating the rest, reducing decimals to a minimum, and storing rotation only on the Y axis. We also keep only the ghosts of cars that complete a race in under 4 minutes to avoid excessively large files.

Because of this, loading 1000 ghosts and then picking a suitable set of 7 close to your skill level would make the process too slow. Currently, we have a maximum load time of 10 seconds for the ghosts, and we are aiming to ensure the game works smoothly even on slow connections.

gus_massa•3mo ago

Have you tried compression?

For example instead of storing, x1, x2, x3, x4, x5, you can store x1, x2-x1, x3-x2, x4-x3, x5-x4 Using few decimals it may compress better with the standard algorithms.

Assuming almost constant speed, you may guess x3 from x2 and x1 and x3=~ 2*x2-x1, so instead of x3, x4, ... you store x3-2\x2+x1, x4-2\x3+x2, and compress it.

For a car, I'd assume constant acceleration and if I didn't make a mistake, x4=~ 3x3-3x2+x1, so you store x4-3x3+x2-x1 and compress the sequence.

andvari_bekho•3mo ago

Hey, thank you so much! I’ll share this information with my programmer, maybe with this we can further reduce the ghost file size and bring in more samples to make the selection better match your skill level.

In any case, the ideal solution would be if the API allowed us to send the player’s average time as a parameter, so we could fetch opponents at a similar skill level, that would greatly improve the gameplay experience. I’ll mention it again to the backend team, and if we’re lucky, maybe they can prioritize that feature.

Cheers, and thanks again!

gus_massa•3mo ago

Too late to edit. I forgot a 3 in

x4-3*x3+x2-x1 --> x4-3*x3+3*x2-x1

Good luck with the project.

PS: Usually tweaking the parameters to select the opponents/ghost and make the game fun (that is what is important) is hard. (More cars? Less cars? Use the last time of the player? Use the average time of the player?) So after you solve the problem, a post-mortem may be an interesting blog post to summit here.

PS2: another typo in my first comment: defectively -> effectively

andvari_bekho•3mo ago

In the future, we plan to reuse the engine and everything we’ve learned from this project to create a new game, this time hosting the backend ourselves to have greater control over the matchmaking.

Right now, we’re limited by what the Gamesnacks API allows us to do, which makes it difficult to fine-tune certain aspects.

Even with those limitations, it’s great to see players enjoying the game. The leaderboard is highly competitive, and I keep seeing the same players battling for the top spot, shaving off just a few milliseconds each day.

I love how accessible HTML5 games are, but I’d also love to see our next racing game running on the Steam Deck. maybe we’ll build a Unity wrapper and release it ad-free on Steam for a few bucks.

andvari_bekho•4mo ago

I have to double check, but I think you can still play without creating a profile if you just close the popup and skip it. However, after completing a race, the game will prompt you again to create a profile in order to save your time.

gus_massa•4mo ago

For the first time in years I'm complaining that it's too hard to notice that I can create a profile. :)

andvari_bekho•4mo ago

Yeah, I think right now the game only asks players to create a profile when they share a challenge link with a friend for a 1vs1 race.

We wanted the racing experience to feel as frictionless as possible, but I agree that for new users it can be unclear that creating a profile is even an option.

From what I understand, Gamesnacks is targeting the game primarily at their existing player base, so they expect most players to already have a profile (since it’s the same one shared across all their games). Still, I think this game has the potential to attract new users to their platform, so making the profile creation option easier to discover would be a good idea.

Thanks for your feedback!

Framing an LLM as a safety researcher changes its language, not its judgement

Are there anyone interested about a creator economy startup

Show HN: Skill Lab – CLI tool for testing and quality scoring agent skills

2003: What is Google's Ultimate Goal? [video]

Roger Ebert Reviews "The Shawshank Redemption"

Busy Months in KDE Linux

Zram as Swap

Green’s Dictionary of Slang - Five hundred years of the vulgar tongue

Nvidia CEO Says AI Capital Spending Is Appropriate, Sustainable

Show HN: StyloShare – privacy-first anonymous file sharing with zero sign-up

Part 1 the Persistent Vault Issue: Your Encryption Strategy Has a Shelf Life

Show HN: Teleop_xr – Modular WebXR solution for bimanual robot teleoperation

The Highest Exam: How the Gaokao Shapes China

Open-source framework for tracking prediction accuracy

India's Sarvan AI LLM launches Indic-language focused models

Show HN: CryptoClaw – open-source AI agent with built-in wallet and DeFi skills

ShowHN: Make OpenClaw respond in Scarlett Johansson’s AI Voice from the Film Her

CReact Version 0.3.0 Released

Show HN: CReact – AI Powered AWS Website Generator

The rocky 1960s origins of online dating (2025)

Show HN: Agent-fetch – Sandboxed HTTP client with SSRF protection for AI agents

Why there is no official statement from Substack about the data leak

Effects of Zepbound on Stool Quality

Show HN: Seedance 2.0 – The Most Powerful AI Video Generator

Ask HN: Do we need "metadata in source code" syntax that LLMs will never delete?

Pentagon cutting ties w/ "woke" Harvard, ending military training & fellowships

Can Quantum-Mechanical Description of Physical Reality Be Considered Complete? [pdf]

Kessler Syndrome Has Started [video]

Complex Heterodynes Explained

MemAlign: Building Better LLM Judges from Human Feedback with Scalable Memory

Framing an LLM as a safety researcher changes its language, not its judgement

Are there anyone interested about a creator economy startup

Show HN: Skill Lab – CLI tool for testing and quality scoring agent skills

2003: What is Google's Ultimate Goal? [video]

Roger Ebert Reviews "The Shawshank Redemption"

Busy Months in KDE Linux

Zram as Swap

Green’s Dictionary of Slang - Five hundred years of the vulgar tongue

Nvidia CEO Says AI Capital Spending Is Appropriate, Sustainable

Show HN: StyloShare – privacy-first anonymous file sharing with zero sign-up

Part 1 the Persistent Vault Issue: Your Encryption Strategy Has a Shelf Life

Show HN: Teleop_xr – Modular WebXR solution for bimanual robot teleoperation

The Highest Exam: How the Gaokao Shapes China

Open-source framework for tracking prediction accuracy

India's Sarvan AI LLM launches Indic-language focused models

Show HN: CryptoClaw – open-source AI agent with built-in wallet and DeFi skills

ShowHN: Make OpenClaw respond in Scarlett Johansson’s AI Voice from the Film Her

CReact Version 0.3.0 Released

Show HN: CReact – AI Powered AWS Website Generator

The rocky 1960s origins of online dating (2025)

Show HN: Agent-fetch – Sandboxed HTTP client with SSRF protection for AI agents

Why there is no official statement from Substack about the data leak

Effects of Zepbound on Stool Quality

Show HN: Seedance 2.0 – The Most Powerful AI Video Generator

Ask HN: Do we need "metadata in source code" syntax that LLMs will never delete?

Pentagon cutting ties w/ "woke" Harvard, ending military training & fellowships

Can Quantum-Mechanical Description of Physical Reality Be Considered Complete? [pdf]

Kessler Syndrome Has Started [video]

Complex Heterodynes Explained

MemAlign: Building Better LLM Judges from Human Feedback with Scalable Memory

Show HN: Ghost Pro Racing 3D Async Multiplayer Arcade Racing in HTML5

Comments