Show HN: I'm building the “chess.com” of speed cubing | Hacker News Reader

Show HN: I'm building the “chess.com” of speed cubing

cubedesk.io

140 pointskashnote70 comments

Open articleView on HN

It would have been nicer if you were building the "lichess.org" of speed cubing

A bit dismissive of the work and pretty impressive level of execution the guy has put in to get to this point.

I personally have quite a strong distaste for websites such as chess.com. I'll just leave it at that.

Why? I much prefer chess.com to lichess. For one the UI and game modes are vastly superior!

Each to their own. I find lichess.org's UI and UX vastly superior.

And I'm not the GP, but I agree. To expand my thoughts:

- The free version is vastly inferior to lichess.org, and it nags you to upgrade at every opportunity.

- They are anti-competitive to the point of censoring the mention of lichess on their forums (for e.g. feature requests) [1]

- I guess it's subjective how paid chess.com compares to lichess.org but I much prefer lichess.

- Minor UX/UI annoyances (you can't give your opponent time or allow takeback requests in a rated game, etc.)

- The price of diamond membership is quite steep for a chess website, but I guess compared to lichess.org anything is quite steep.

Thinking about it further, I guess chess.com seems extra bad because of the existence of lichess.org which is just amazing (100% open source including the apps, zero ads, open API, can download the database of all games, etc.)

Full disclosure, I'm a diamond chess.com member (signed up in January for a year) because I'm quite a chess enthusiast and really wanted to see if it was worth it. It's not, in my opinion.

[1] https://www.reddit.com/r/chess/comments/qudyuq/why_streams_w...

FWIW, I'm happy to sponsor both, they innovate in different ways. Chess.com with their content sponsorship and deep player pools, lichess with ad-free, open source and unique UI features.

Yeah, I strongly dislike chess.com, but the fact that it sponsors so many amazing chess content creators makes it more useful existing than not existing. Too bad most players gravitate towards it over the superior platform (lichess).

Yeah, in an ideal world lichess would get so many donations they'd buy chess.com, make it free, and sponsor the same content creators :)

It's not just content creators. They're involved in events, marketing, and all other sorts of stuff that's dramatically increased the reach of chess. You really can't overstate how much they've done for chess.

I've played online since back in the days of ICC which was a defacto monopoly given how much larger than it was than all other sites. You can see their records here [1]. Their record for the max number of players logged in was 7,352. Chess.com at a really random time right now has 183,538 people playing.

It wasn't long ago that all of the big tournaments were gradually dying off - even Linares died, for lack of funding. And then the rise of chess.com (along with Magnus - who is an absolutely amazing world champ in terms not only of play but also 'ambassadorship for the game') just sent the game in an entirely different trajectory that nobody would have imagined just 10 or 15 years ago.

[1] - https://www.chessclub.com/help/records

> You really can't overstate how much they've done for chess.

Yes, it's great that this game is finally getting popular. About time an up and comer took the crown away from Settlers.

It’s become somewhat cool and modern again with chess.com. It’s hard to ignore that. I quit playing competitively around when I turned 14 because it was so unpopular/geeky for someone that age, but with chess.com pushing marketing and streaming so successfully it seems much more likely a teen will carry on playing.

I think it was more The Queen's Gambit.

Interesting! UI and app are way better on chess.com in my opinion. Game analysis is heaps better on chess.com (UI wise, content wise they're equal).

The money nagging is annoying, especially considering they're charging for basically just stockfish wasm running on your machine.

I stopped using chess.com because I found it way more addicting compared to lichess and I was way too invested in my chess.com rating.

Here is Magnus Carlsen describing chess.com which sums up how a lot of us feel: https://www.youtube.com/watch?v=brrFMwNVWFA&lc=UgzmCWEI-gfzv...

lichess.com does have a snappier feel. Using the chrome performance tools it seems that lichess takes about 2ms to go from mouse click to the piece selection repaint and lichess takes 12ms. (screenshots of traces: https://imgur.com/a/iUkGm8F)

Lichess also uses zero-time premoves and chess.com uses .1 second premoves. I think these also contribute to lichess feeling snappier.

> Lichess also uses zero-time premoves and chess.com uses .1 second premoves. I think these also contribute to lichess feeling snappier.

Having played competitively, zero-time premoves make no sense and they should cost.

Andrew Tang would probably disagree

Probably and I was nowhere near that level, I just played for my county and tried for my country as a kid, but it changes blitz massively and should be it’s own game. It’s not a thing you can do IRL, so unsure why it should be a thing in the online representation of that.

For the past couple of months he's been exclusively on chess.com, with the exception of exactly 1 15 second game he played on Lichess, so I think his position might also be a bit more nuanced!

Lichess is great if enjoy the flagging meta, and not so great if you don't. It's made even more extreme by their interpretation of sufficient mating material. A single knight is generally counted as sufficient mating material so you end up scenarios where you're forced to play N v N along with 0 second premoves and that's not something I much enjoy.

I understand some folks like this sort of stuff, which is why it's great that there are multiple sites competing instead of a single site monopoly.

Please don't just leave it at that. As you expressed a strong sentiment publicly, I'd like to why.

(not OP) lichess is free software, it's free software that works, it's community supported and sustainable. chess.com is a mild piece of software and a company that splits its time between trying to be a business and trying to provide a good chess experience. I'd prefer more of the former to exist.

I like the lichess story a lot. Last time I read an interview with the guy running it, and I really appreciated his attitude towards his service and to what it represents to people.

The point is that it’s not a compliment to call something the “chess.com”.

lichess is much nicer to use. Try starting a chess game with someone who's never used it before on each and see the difference.

I came here to comment that it'd be better to build lichess of speed cubing (whatever this is), but you beat me to it.

Speed cubing? Let me google that for you - https://en.wikipedia.org/wiki/Speedcubing

Better build the lichess.org of speed cubing

If you don't like money

If you don't like money more than freedom and honour

at this point it's more than just money.

lichess is my go example for what you can do with community driven open source software.

I've always loved this clip of Lucas Etter posting a speed cubing record time, in what appears to be a gathering of Stranger Things kids in real life...

https://youtube.com/watch?v=vh0W8E4cNkQ

This looks great! I think it's going to blow up in a massive way if you can get livestreamers on Twitch and YouTube to check it out, so be sure to try to reach out to those.

I think cubing definitely has potential to become mainstream like chess has. Maybe I should try reaching out to chess streamers to share my app with their audience.

Ludwig Ahgren would probably be a really good fit, I think he’d love the opportunity. :)

And he's good friends with much of Offline TV, a chunk of their extended circle are accomplished cubers (Edison and Leslie being the standouts)

Maybe I'm the wrong target audience. I spent some time last summer figuring out 2x2 and 3x3 from scratch, without referring to any existing literature, methods, etc.

That worked but solving a 3x3 is in the 2-3 minute range. FAR from speed cubing.

OK. Now I'm interested in getting into speed cubing. The Trainer shows like 12 systems with no descriptions/trade-offs/recommendations ... how they are related ... best order to learn them if learning several... etc. Those points would be SUPER handy.

Anyway... The rest looks good and I'll definitely play with it.

You're right, it's not really beginner friendly. I want to improve upon that in the coming months. But for now, there are tons of cubing-related resources online nowadays. For 99% of cubers, you want to focus on OLL and PLL algorithms.

It's mostly memorising.

I memorised a few basic algorithms a few years ago and practiced getting faster.

Then you basically add algorithms to take shortcuts and improve your hand speed.

I never got good times (maybe around 1m?) but it gradually lost appeal after a while. Nowadays I probably would get stuck trying to fix the last 4 corners.

I'd like to get back into it one day and learn some very generic algorithms to just be able to transition between states, without the explicit goal of solving the cube fast.

There is nothing missing from this timer, it's very impressive! You might be calculating larger averages slightly incorrectly? Convention is discarding best and worst `ceil(0.05 * len(n))` times. I've seen a lot of timers come and go what will be different about the longevity of cubedesk?

I'll double check the large average calculation logic.

Good question. I recently came out with a Pro version for CubeDesk which is mainly there for people who want to support development. You also get some minor stuff like additional themes and 1v1 options.

I'm hoping that enough people buy Pro so that I can cover the server costs. As long as the server costs are covered (or at least not too high), there's no reason for me to ever take down the app. If nothing else, it's a good learning opportunity for me.

What makes this different from similar sites that already exist like cubingtime?

So this has no interaction with the cube itself? It's just a stopwatch app that you press the spacebar to stop, and everyone trusts that you didn't cheat? It doesn't have a programmable hardware LED cube that the app dynamically loads problems onto for each round and logs your moves? Short of that, with a normal cube, it doesn't make you hold up the sides of the cube to the camera before the round to let it determine the problem? Nor make you hold up the cube to the camera after completion to ensure that you solved?

Your camera-based solution would be trivial to cheat at though; simply have two cubes where one is in the solved state. So, incredibly time consuming to implement while also not really helping much aside from security theater.

Right, it's a hard problem to solve. A camera-based approach would be invasive, impractical, and really easily duped. But now we have bluetooth cubes.

Bluetooth cubes still have a ways to go before they're physically good enough and also affordable enough to replace regular cubes, but they will solve this issue. CubeDesk already supports a few types of BT cubes. When you use a BT cube on CubeDesk, you get a breakdown of every stage, turns per second, areas you paused, etc.

The unfortunate thing about BT cubes is that all the companies making them have their own proprietary software that they want their customers to use, so they encode the data coming from the cube to make it really hard for anyone to make sense of it. Some of us in the cubing dev world have reversed engineered these techniques, but it's just unfortunate how they try to keep it a secret.

It looks like it's secret. I couldn't find in GitHub anything related with reversed engineered bluetooth cubes.

My first reaction glancing at the site was "Bluetooth Rubik's Cube". I could confirm in seconds that such cubes exist, but minutes spent spelunking the site gave no evidence that CubeDesk had even considered this issue.

Then I opened the answer key (HN comments) and found this thread. Shouldn't you make it easier to find documentation before signing up?

I'm working on it. I recently redesigned the landing page and plan on adding more info soon. I don't work on CubeDesk full time so it's mostly after work or weekends when I can make improvements. I'd like to add a whole help/docs section soon with all this info.

The leaderboard / ranked function will be literally useless. I don't see at all how you're supposed to prevent cheating here.

Did you design the landing page yourself? It looks cool :D.

I'm also curious if you used a tool like Webflow for it.

Thanks! Yup it’s all made with React and Sass.

Didn’t use anything for designing it. Just mainly eyeballing

Why do I have to sign up to do anything? I'd love to re-learn some algorithms, but forcing me to sign up to access anything at all just makes me want to not sign up. chess.com lets you do a bunch of stuff without signing up.

On chess.com everything is locked except for the ability to play a guest game. Even by signing up you hit a paywall at every turn.

I actually have plans of making some parts of the app accessible without having to log in.

Since I started the app by saving data to the DB instead of the user’s device, this is more of a technical limitation than a choice that I’ve made. I just need to put aside the time to implement it.

Super cool. My 11yo son has become a speed cuber. Will point him this way.

really nice. I'm trying to be speed cuber. currently around 40-50 secs for solving.

Did you defined ranking program? what is the default rank? is it going to be similar to chess (e.g 1000 for starters, 1800 for experiences, 2300 for masters etc) ? how are you going to measure it? would it be changed only when you do 1x1 games? or also self training impacts it?

also, how can you protect against trolls? everyone can fake it, you know.

Checkout https://www.worldcubeassociation.org/

I have implemented an ELO system, but it's just running in the background right now with no UI component. I'm collecting data for now to see how I'll need to tweak the various params that go into ELO calculation. I'm going to try to make it as close to chess as possible.

As for cheaters/trolls, I have a manual reporting system right now, but I'll do my best to start flagging accounts automatically. Pretty hard thing to solve given the nature of the issue.

This is awesome. Just watched some Youtube videos of people using the app, looks great.

Is it only for the 3x3x3 cube?

To add to that, is there a good tutorial for 4x4 and 5x5?

Might be good to point to tutorials you have tried and why you didn’t find them to be good. Otherwise it turns into - let me google that for you - here https://jperm.net/4x4

Jperm is generally pretty good but didn't explain all the algorithms, when I learned a 4x4x4 a couple months ago I used that plus this http://www.rubiksplace.com/cubes/4x4/

Basic gist is solve the middle 2/3 edges to look as if they were a 3x3 edge, then solve like a 3x3, accounting for parity issues. Beginner method at least

It can be used for any cube or event. The WCA ones are built-in but you can even add custom event types if you want.

Very cool!!

great design!

I know that the HN is not fan of crypto but I would like to add a decentralization point of speed cubing: "How To Model A Rubik’s Cube in Algorand’s TEAL" [1] this was written by Guido who is one of the top speedcubers in Latam and was invited to join the discussion here.

Algorand was also selected for official chess ratings [2].

[1] https://blog.coinfabrik.com/uncategorized/how-to-model-a-rub...

[2] https://www.algorand.com/resources/ecosystem-announcements/a...

This has absolutely nothing to do with speedcubing apart from describing an algorithm and data structure to save speed cubing in blockchain.

It is open source code that could help other projects to build and mesh upon. Keep the imagination going please!

If it is a open source code that uses blockchain for the sake of blockchain without actually providing anything new, no.

Reply on news.ycombinator.com