Using AI To Make Web Game Hits like Sumplete

Hello! Can you tell us who you are and what you do?

Hi! I’m Daniel Tait, the creator of games you see here on the Hey, Good Game website! Those are Wordga, Crosswordle, Sumplete and Mathler. I’ve also been working on something new recently called wordnerd.co, but today I’ll be talking more about Mathler though!

Mathler is a Math-based game inspired by Wordle. It requires you to use basic Math operations like addition and multiplication and the digits zero to nine to create an equation that equals the answer given. There are multiple modes of difficulty with different conditions so everyone can play at a pace that suits them!

I’m happy to be here to share my stuff with you all!

What’s your favorite game to play?

My favorite game. That is a very difficult question that I probably should have a better answer for since I knew the question was coming. I would say recently my partner and I got a Nintendo Switch and it's the first console that I've owned in probably over 10 years now. So it's really nice to get back into some proper console gaming. We have a massive backlog of games that we've been wanting to play over the years, but just never had the means to. So we're just working our way through those.

Right now we're in Zelda, Breath of the Wild and we're playing it in a sort of couch co-op style where we just take turns with the two of us. So it's been quite a nice evening activity for us when we check out from work, but I'm really enjoying it so far. It's an awesome game.

A family of puzzlers

I come from a family of puzzlers. For puzzle games specifically, my mom plays every night, probably two to three hours of puzzle games, a whole mix of things from the Crossword, Sudokus, jigsaw puzzles and all sorts of things. And my dad, on the other hand, he's a bit more of an interesting story. So he's been playing Scrabble on the original Nintendo DS. Scrabble 2008 is the only game my dad owns and pretty much the only game my dad plays. So every night he'll sit down with his DS for around an hour while they're watching TV together and just play away on that almost 20-year-old game now. So a simple man with simple pleasures, that's definitely his go-to.

How did you get into game creation?

I used to tinker quite a lot with little games way back in the day, even back in high school and in my early university days. I was the type of person who was right-clicking files on the PC when I was 10 years old and trying to figure out what was inside them. And that was pretty much how I got into these things. I think the first game I ever remember tearing apart was a game called Tunnel Blaster on the PC, and I haven't played it nor looked it up in a long, long time, but it was a very simple PC game back in the 90s. 

I just tore apart the packages inside. I found the puzzle files with the graphics and all sorts of sprites and things like that. And I started just tweaking them and making my own levels for that game, which was the first time I started thinking, “Okay, this is interesting. I can change this here and I can change that there. And I can change the speed of the enemies here.”

And so that's kind of like my origin. I've always been a big gamer, so I guess it just brings a natural curiosity to know how things are made. But to sum up, I didn't really spend too much time in game development. Over the past five to seven years I've tinkered around here and there, but nothing until Mathler was ever released or shown to the world. So yeah, it was nice to finally put something out there.

Why did you create Mathler?

I sadly don't remember the details that well. I'm the type of person who is kind of always hacking on something at the weekend. And most weekends I'm working on side projects and all sorts of things you can see that I've released. Then for Mathler specifically, Wordle was taking off massively at the time. I was playing daily, my whole family was playing daily and we'd already been playing for months I think, before I ever thought of the idea for Mathler.

And I think I was just sitting there at the computer one Saturday wanting to build something. I thought, “Okay, let's play around with Wordle and see what we can come up with in terms of some variations or some tweaks to it.”

Then how I actually went about building Mathler, I found an open source Wordle clone online. It was built on the rails of a very well-put-together Wordle clone that someone released on GitHub and I just started tinkering with that file. For the first version of Mathler, I think the only change that was made was I converted that word list into a list of mathematical equations. There were no other significant changes made to the open-source Wordle game. And I just thought, “Okay, this is interesting. Let's give that a go.”

And yeah, it was fun to play. I'm not even into Math really, but I enjoyed the game. It's not a category that I've ever really spent much time thinking about, but it seemed like an obvious variation to try. And it just seemed to work. It was just one of those sort of serendipitous lucky moments that it worked straight at the bat. And I think I released it that day on Hacker News. I posted it and it started to get some traction. And yeah, I had to go through some development after that. But that was that was the origin of the original idea.

How did you promote Mathler?

So for Mathler, it was solely a single Hacker News post that launched it. I didn't even post it on Reddit or any other forms at the time. And if we were to look back, I don't even think it was a huge successful post on Hacker News. I think it got about 70 points or something, so it wasn’t like a massive post. I think it was just the Wordle hype at the time was so high. And then people were sharing scores in groups. And from that point, it really grew up quite fast organically with just a ton of direct traffic. So I've never really been able to source or pinpoint the exact moment that it started taking off. But that was the only remarkable thing I did for it.

I didn't reach out to anyone, but I think Tom’s Guide is the first website to do a write-up. And from there, I think most other publications just took the information or the story from there when they were writing sort of different Wordle variations that were out there.

What inspiration did you have in mind when you made your games?

When you have something successful like Mathler and then you start to have an audience, it just felt like I was such an indie hacker. I mean, I like to make these side projects that just felt like such an awesome opportunity to build more things for an audience that was already there. It's just such a privilege because I'm very used to releasing things that get 10 eyeballs and then that's gone into the abyss for the rest of time, you know? So it was really nice to have an audience. 

I really took full advantage of that and just started churning out games at quite a high pace. So I think probably since then, my folders must have 50+ games in there. I've tried to throw some of them out there and see all sorts of variations on common classics and then some more creative endeavors. 

You guys may have seen quite a few of them. There are some that have done much better than others in that group, but it's been great fun. It's just amazing to now launch something and have 1000 people maybe look at it pretty quickly. And that's just a nice loop, you know, to start building more things or to change features. And it's something I had never had before. So it was a really nice change.

I play a lot of games which obviously helps for inspiration. Like I play tons of whether it's puzzle games, larger titles, just all sorts of games. So I think I understand basic game mechanics quite well in terms of what core a game has to have, like a key feature or a mechanism. But I actually then come up with the different ideas. I have my design process, and I'm a weekend builder because I have a full-time job during the week as well. So I typically set aside my time and I also kind of set limitations for myself.

So I say, “Okay, I've maybe got six hours to work on something. What do I want to build in that time?” And let's say I'm interested in building something in the word game category today. I have a list of ideas—very one-level ideas. When I'm on airplanes, I typically sit there staring blankly, no Apple, not on my iPhone, just sort of scribbling down very rough ideas for games or other websites and things like that. So I've got a long list of things that I want to build at some point.

And then I set aside the time to throw up something as quickly as possible to the point where I can play it if it's a game or I can use it if it's a website. And I just see if there's any sort of spark there to try and develop further.

A “throw and see” approach to games

I think it is important to sometimes just realize that things aren't working or your audience don't like what you've created. It's perfectly okay for that to be the case. You know, I've made some terrible, terrible games. I'm happy to admit to that. There are many games I play for five days and I don't enjoy them at all afterwards. So it's also a self-learning process that I may do differently than some other creators. 

Rather than playing the game privately, I'm going to that testing loop in my own sort of local space. I just throw it out online and get some other people to test it alongside me. So there are many times I launch a side project or a game and then a few days later or a few weeks later, it just won't exist anymore because either I just don't think it works or the audience has kind of told me that already. 

So yeah, I'm a big fan of building on the web as well. It's so easy to do that and there's a really quick feedback loop. I don't need to go through any approval system or any build process to launch things. So it's just easier just to throw it out there.

Can you tell us more about the features you added to Mathler?

After its launch, the difficulty levels were added to it pretty quickly once Mathler had reached a peak or a very large audience. At the time, I thought, “Okay, well, rather than launching new things, I should really focus on finishing or polishing it.” It's one of my more negative traits. I am not really a finisher and I'm definitely not a perfectionist, that's why you get a portfolio of 50 games rather than one really good game. 

But with Mathler, I felt a deep responsibility to treat it well and to actually develop the game further when it had such a large audience, and I had been receiving a ton of positive feedback from people asking to develop it further. And I think some of the ideas such as the different difficulties actually came from the the little feedback button on the the website. I think some people mentioned we can try this, we can try that. I got tons of inbound those first few weeks from people with ideas. 

But I think some of the ideas for the difficulties definitely came from the community, so I went ahead and built those. The funniest part of the story of Mathler is that the tech stack from Mathler was a stack that I was completely unfamiliar with, to the point that I had no idea how to edit the game very well. It was built in this TypeScript React tech stack and I just have never used it before. I was hosting it in Netlify, which I'd never used before. I didn't know how to run a build through it. I didn't know how to do anything.

So when I was making changes, I was actually editing the game directly in the GitHub web code editor, just writing changes and doing commits every sort of 15 minutes or so to push changes live. So it was a very, very dirty and messy process to make changes to Mathlerr at that time, but I kind of just kept going with it until I got features out. And yeah, that's where the different difficulties came from, as well as some of the other features that came later.

On not truly identifying as a traditional “programmer”

For sure, no. And I would say that obviously I have some programming skills for sure. And I understand some of the basics behind programming. But I'm just not a programmer by trade. And I'm sure when a lot of people look at my code, they're like, “Oh my goodness, what is happening here?!” Because some of the things I do, I know myself that I shouldn't be doing them, but I just like to build things. I don't really mind if it's design, coding or writing, I just use the tools that are sort of at hand to get things made that I can play with. it's kind of what I love to do.

On how he views himself as a professional

Nowadays I have no idea how to introduce myself to people publicly when work is brought up. Honestly, I try to avoid the question. I just usually tell people that I just work on the computer and I usually don't get too many follow-up questions. At least in the circles around where I live, that's usually a conversation-ender. So it's quite simple for me to just try to move on from there.

Did you monetize Mathler after the launch?

Yeah. I can't remember exactly, but I think it was quite a while. Maybe two or three months after the launch, I did put ads in it. I think monetization is important and I would love to when I have a large audience, I think it's important to capture some value from that site so you can spend more time on them and you can improve them and you can release new products and features to that community.

I think it's sometimes a mistake to not monetize these products as you can give it your all when you have the opportunity. But I've been saying that I think a lot of products do monetize a bit too early and it shouldn't be their priority. I personally don't get a huge negative response when, for example, I add ads to a game, but I do find that the engagement rate goes down a little bit in terms of the community size. I feel like those people who are reaching out to you with some ideas or some product feedback just start to decrease. That might also just be the life cycle of the products. So it might be nothing related to the ads. It might just be people have already shared what they wanted to share with me, but I have definitely seen that in the past.

So I wait until I'm happy. I don't typically monetize until I get the products to a point where I'm actually not planning to change it too much after that. Although that's more relevant to more simple products like web games where I typically get them to a point I'm happy with and then don't develop them too much further. And yeah, monetization in the past has been simply through just display ads. I haven't really tried any other forms. I would love to try some other ideas and how to monetize, but it just hasn't been a priority so far.

Could you tell us about Sumplete?

For sure! Sumplete was made almost entirely through ChatGPT with one of the early versions of ChatGPT, I think it was maybe the 3.5 model. It was the first time I was testing these new AI tools to see what I could do. You could probably tell from some of the stories I've told earlier that a tool like ChatGPT, for me, is an absolute gold mine for someone who just likes to build things quickly. And to have a tool like that on my side to help me with some of the more technical challenges. It's really unlocked a whole new world of side projects for me, so many things that I would never been able to build in the past.

I can now get some help from the AI assistant and yeah, build something that may be a bit more complicated or a bit more interesting. So Sumplete was partly that. So again, this is going back in time a bit, but I think the origin of Sumplete with ChatGPT was when I was essentially talking to it and asking it for ideas for a puzzle game for me to make. It wasn't my original idea, wasn't actually doing any coding with ChatGPT. I was just using it as more of an idea sounding board to say, “Okay, have you got…”

And then my original question to ChatGPT was “I really like playing Sudoku. Do you have an idea for a similar style puzzle as Sudoku?” And I came back with a few different results, a few different ideas. And the one that stood out to me was what turned out to be Sumplete a day or two later. So it was essentially a number elimination game. I think some people described it as a reverse Sudoku. It's a little bit like that. There's obviously logic involved and a bit of Math involved as well. 

It was definitely an interesting concept from the first time I saw it. What was particularly interesting was that I could play it inside ChatGPT because it was able to give me a grid that I could actually play with and try to solve within the chat interface. So straight away I had an MVP where I could play a few rounds of the game without actually writing anything. So it was a great way to test whether it was fun or not, and it was an interesting concept and I personally really enjoyed it. Number games weren’t even my sort of specialty!

So I thought, “Great! Let's develop this a little bit further.” And that's when I thought, “Well if we've come this far, let's just see how far we can push it.” So I made follow-up questions asking ChatGPT to actually write the first version in HTML and it got me a playable version online. And then from there, I started working with it to create a generator so it could generate unique grids with unique solutions.

And that, for example, is something that I would never been able to code myself, that sort of algorithmically generated puzzles with unique solutions would be beyond my capabilities for other projects. So that was a huge unlock for me. And yeah, it was a really interesting project to test these new tools. The audience received it really well. I think the story really was an interesting point for people, and it was probably one of the first web games ChatGPT helped build. And it got lots of positive and negative feedback for that because there's a whole range of different opinions on whether that's a good or bad thing.

So you use a lot of AI with your game creation?

Yeah. So I went through a phase of maybe six months where I was using it a lot to generate all sorts of things. I've actually kind of reverted back a little bit now and I'm doing some of the more simple stuff just by myself. I kind of missed actually writing some code or I'm doing some things on my own, but I still use ChatGPT a lot of the time to check things and to give me a second opinion, especially on my programming to tell me why things aren't working. I was the type of person who was spending hours and hours on Stack Overflow, so it really replaced that sort of need for me. And it's been a huge unlock for me.

What keeps you building games?

It's just really nice to build something. So games in particular are things that I've only been building since Mathler. So before Mathler, I was typically working on side projects for other stuff. I made an online birthday card website for a while, I had a website where travelers could meet each other in different cities and so on. I really like building things that people can use.

In my other full-time career, I built B2B software, and I work for a B2B software company. So it's mostly doing work for consumers and products that I don't use myself or my friends and family will never use. It's just really nice to build something that I can send to my family group chat and say, “Oh, hey, guys, I made this. You want to give it a go?” And I know they love puzzle games and they're like my number one fans when I send them puzzles. 

So it's really rewarding to just build something for friends and family to use. And yeah, it's just pure fun, isn't it? Just starting from a blank page and coming up with something then putting it out there. It's amazing. We can do it from this little machine in front of us and it’s great.

On doing side projects alongside his day job

They're very almost polar opposite pursuits, I guess. One of them is trying to solve a very specific problem for our large companies, in my case, regulation and reporting. So not the most exciting field, but something that's needed. And they're useful for those companies, but it's not a product that you get a ton of enjoyment from the design process and you never have thousands of people using it. It's typically targeted to a single user. And it's to sort of replace an old Excel spreadsheet somewhere with a slightly more useful tool.

Whereas for games, I just find that I love them because it’s partly just my own hobbies and passions. I play my own games every day, and I'm not afraid to admit that, however embarrassing it is. 

Me and my partner sit down together and we kind of go through them and my family plays them as well. So it's just very different than I guess sort of serving a customer. I feel like you get to be a lot more creative and build something for yourself, which you also know, maybe some of your friends and family will enjoy as well. So yeah, very different process. Both can be rewarding for sure, but I think they tickle different itches, you know?

Can you tell us about Crosswordle?

There were a few games that had come out before Crosswordle that had a similar idea in terms of mechanics. So you have Waffle, for example, which is the 5x5 grid with the five-letter words where you have to swap the letters to solve the puzzle. We loved Waffle, me and my family. We used to play that most days. I guess the biggest change was we thought this would be really interesting with more variation of word length and more like a crossword-style layout. 

So I had another game already that I previously built called Killer Crossword, a game you guys probably aren't familiar with, but it's actually been online the whole time. I just don't really talk about it, but it's a crossword website that I launched a while ago. And because of that website, I had a crossword generator already built that generates unique crosswords unload. So I just took the code from that and plugged in the coloring dynamics from Wordle and then Crosswordle was born pretty quickly after that.

So it was actually a really quick game for me to build. And yeah, we love it as well. It's one that I still play and it's challenging. It's a hard game. I find it quite difficult and I enjoy the challenge of that one.

On Killer Crossword

Killer Crossword is a funny one. That was one that had a big spike when I launched it and it kind of just faded away and I ignored it for too long, I think. It runs completely by itself. So I just never look at it anymore. It just kind of ticks away. And I occasionally get an angry email from someone because I have a British spelling in the crossword or something. 

What would you have done differently if you had the chance?

The products are businesses I'd sold. So it was a new experience for me, that whole process. It was a surprisingly easy process, definitely easier than I was expecting, which was a nice surprise. But yeah, it came with definitely pros and cons.

Obviously, like I guess one of the pros straight off the bat is that it was great for me financially. It gave me a bit of a security blanket, I guess for my other business and just my personal situation obviously improved a bit once I sold the games. And it just gave me a bit of stability that I hadn't had in many years as a startup founder, as you can imagine. So that was a big plus. 

The second plus was probably there was a weight lifted off the shoulders. Although I loved having the portfolio of games, it was also a burden to know that you have an audience with expectations of you improving those games or maintaining those games. 

And when I was doing it part-time, I felt like maybe too much of my time was going into that and maybe some things, some more important things like friends and family and my partner were starting to fall off the wagon a bit too hard. So I think Hey, Good Game saved some important friendships for me. So that was a big relief. 

The negative side is obviously I lost a big part of my audience because they were within the games themselves. So that was definitely felt when I released my next game and I couldn't post it on Mathler anymore and get that sweet referral traffic. So that was a big mess, but really, no real regrets.

On the process of launching a game now

Yeah, so I had a reasonably successful launch recently. So I had a game called 23 Words. The idea of the game is it's a word jumble that you have to solve, and it’s 23 of them in a row and you get 30 seconds to solve the jumble. If you can't solve it, then you lose. So maybe you solve 13 that day, or you solve 20 out of 23 that day, you've already done really well. 

That idea had a whole sort of different stages of evolution. So the original idea was an unlimited word jumble where you just go through jumbles continuously and you get a streak and a high score. And I built the first version of that on a flight, which just sat there. And I just coded a basic version of that. And me and my girlfriend sat on the plane and played it for about 20 minutes and concluded that it was fun. It's a classic word game style. But what can we do to make a twist on this? 

So then I was just creating different variations. It was like, “Okay, I think we need to have a lose condition for this game because it's fun enough to play continuously, but it does get a bit repetitive.” and of course boredom sort of kicked in and I think I lost interest pretty quickly. So how can I get a bit more interest? We need to have a fail state. So then that's when we added the timer and then when you can lose after a few goals, that makes it a bit more interesting because you can start to compete with yourself. 

You’d be like, “Okay we got five this time. We got six this time.” But I still felt like when I loaded the game, it was like, “Okay, I'm going to maybe play this once or twice and it was fun, but I'm not going to really come back to it.” 

So then the next stage was coming up with a preset target before you start the game. And that's when the idea of 23 words came in and say, “Okay, we're going to set a target. There are 23 jumbles to solve. How many of those 23 can you solve?” Why 23? That goes to the issue of domain name availability. I just went through all the numbers from 15 to 30 and saw which domain names were available and 23words.com was available. So 23 words it was. 

I get that occasional email at least twice a day on why it's 23. And I just keep talking about the same answer. It may not be a very exciting answer, but that's the reason. So 23 Words and yeah, it's doing really well. 

I then went ahead to do my normal launch strategy, which is just a very simple Hacker News post. I've started to post on Reddit a bit as well. And it got received really well by that audience across from a couple of my existing games.

Quick launches through repetition

So I've just gotten over launch problems and development through repetition. I feel that I can release a lot more polished products now. But I just think my products have gone massively downhill over time in terms of aesthetics and because of the reason that I just released them quickly.

I look back at some of my original posts and I'm like, “Wow, did I build this? I don't even remember being able to polish something out that nice.” But yeah, after spending so long on projects that just gained no traction, I just started thinking, “Well, what's the point in this? Let's just release things a bit quicker.” And I started to just throw things out and try and get it over with.

I guess one of the techniques for that is I don't really ever read the comments, which is not great in terms of feedback in the Hacker News post. But I let the feedback come to me via the contact form on the website because I find that the feedback is far more constructive if someone opts in to sort of send you back, whereas Hacker News is not the friendliest place at times. So I skim through them, but I just avoid any of those negative comments saying “This is a waste of time” or Bah, I could make this and even better in 10 minutes!” and so on. So better to just ignore those, I think.

Is there anything you’d like to promote?

Yes, Word Nerd. The easiest way to describe is obviously for people to go and check it out and have a look. But in short, it's a collection of all my favorite word game ideas that I've had over the years in my little notepad. So right now on the site, I've got five games live, and I've got probably three more coming in the next few weeks to sort of complete that set. And I think it will stay there for a while. I think that's going to be sort of the cap.

But the idea is that I don't feel like word games have like a true home right now compared to some of the other categories. I feel like there's, as a word nerd myself, no default place that I go to play games. The closest one would be the NYT for sure. I enjoy their crosswords and crossword is such a core game in the word game category. It really saw a lot of other word games circle around the original crossword, but some would argue crosswords is actually quite a lot of trivia and crosswords rather than sort of word semantics and that sort of more word puzzle logic.

So I enjoy more of that side of word games; word semantics more than the trivia. So a lot of my games that you'll find on Word Nerd are a lot more to word construction or finding words within words, etcetera. So a different style of word game I'd say. And yeah, the best way is just to give them a play and see what you think.

Final advice

I was thinking about the one thing I would want to actually say, and I think it really is just to tell people to go and build things. I think that's what it's all about. Stop overthinking it and overplanning and worrying too much about what people think. Just build something and throw it out there.

And I want to say I have the most fun in the world just building some terrible, terrible games that are forgotten about a week later. But eventually, you'll find something that you actually like to play every day yourself. So just keep at it and enjoy it.

Where can we find you to learn more about you and your projects?

I don't have much of an online presence really so I would just refer people to Word Nerd for my current project. You'll find a contact form in there in the menu and you can drop me an email. I'm quite I'm old school, so feel free to anyone to drop me a message. I'd love to hear from other game developers or just anyone interested in getting into the spaces. I'm always happy to have a chat.