Say you want to build an online web-based game of (say) Hangman where the person who continues to win eventually wins a cash prize. The more players that enter the game the more the eventual winner wins.

However, it is clear these days, that any game which has cash prizes will attract the bot-writers. I would assume that even a game like Hangman would be a trivial matter for an experienced bot-writer to create a bot which would provide a significant edge over a human player.

Based on this fact; that any online game could provide a bot-writer with a advantage over human players, my question is: is it possible to create a online game which provides cash prizes which a bot-writer would have no significant advantage over human players? Or is it simply that bots are too powerful and pervasive now in gaming and a real-money-transaction game is not feasible?


4 Answers 4


There are four kinds of games I could imagine where human players wouldn't get outpaced by bots.

  1. Embrace the bots. Make a game about bot programming. Let the best bot-writer win. It's not cheating when it's the point of the game :) (Ok, this is kind of a cop-out)
  2. Create a very complex strategy game with large numbers of units to control, a large number of interlocking mechanics and many, many decisions to make (like Civilization). The standard MiniMax algorithm which beats human players at games like chess can't be used here because the decision tree is just too wide. I have yet to see a 4X game where the AI can pose a challenge to a moderately experienced human player without having to resort to some form of cheating. Turn-based games might be better for this than real-time games, because in real-time games AI opponents can benefit from their superhuman speed and perception when it comes to micro-management.
  3. Make a social deduction game like Town of Salem where players need to communicate with each other and win the game through psychology, diplomacy and deception. Bots will fail the Turing test and get ganged on by the human players. Most social deduction games I know have some design flaws which prevent them from being really e-sport compatible, like imbalanced roles or bad players having it too easy to screw up the game for good players. But I am curious how a social deduction game specifically designed for competitive play could work.
  4. Make the game a popularity contest. Give the players a creative task and then have humans rate them based on aesthetics and originality. There are some pitfalls here, of course. When you design the voting system badly, it might be vulnerable to brigading, tactical voting or bot voting.

Your game has to have some form of income to make it work - it has to make money so you can give money away.

Assuming your goal is to make YOU some money...

Base your game concept around the idea that even if a bot is playing, it's still making you money. More money than the bots are being paid. Done right, you only care about bots upsetting other human players, because they aren't costing you anything (beyond maybe human players).

  • \$\begingroup\$ But why would any human want to play a game where there is no chance of winning the prize due to bots? Then just the bots will be playing against bots. That also opens up liability if a human player feels that bots are stopping him from winning a cash prize and the game developers aren't doing enough to guard against bots. \$\endgroup\$ Jan 2, 2019 at 4:49
  • \$\begingroup\$ Like I said, are you out to make money or make people happy? My answer is the "out to make money" answer. The "make humans happy" answer is really not possible without knowing what your game is about, the gameplay concepts, etc. \$\endgroup\$
    – Tim Holt
    Jan 2, 2019 at 4:50
  • \$\begingroup\$ @ImTalkingCode Even if your game ends up as "bots vs bots", then only the best bots wins, which would make your users happy, as they apparently enjoy making bots. Assuming there's an entrance fee, you still earn money. \$\endgroup\$
    – TomTsagk
    Jan 2, 2019 at 10:46
  • \$\begingroup\$ @ImTalkingCode another option would be to make lots and lots small and simple minigames (maybe procedually generated) and then make the player solve the minigame. This way bot writers will find it harder to write a bot that can solve every minigame, plus playing the game too. Another crazy idea would be to randomize the position of the UI in order to not make bots work XD \$\endgroup\$
    – user115399
    Jan 2, 2019 at 13:19
  • \$\begingroup\$ Consider gambling, casinos, and daily fantasy sports. The vast majority of players almost always lose money. A few sharks may turn a profit regularly, but only the house turns a profit 100% of the time. Most gamblers know this but plenty of these low level players show up again and again anyways. These examples serve as a rough proof of this type of business model but you need to assess the specific ethical/moral context because this is a place where the line is blurred between gaming and gambling (same goes for loot boxes). There is a thin line between engaging game design, and a Skinner box. \$\endgroup\$ Feb 14, 2020 at 19:27

Come up with a game that is too hard to write a bot for. For the vast majority of games out there, the investment of writing a bot is vastly higher than the prize you could get for winning it. For most professional e-sports, bots simply aren't good enough to compete with capable humans. (We're getting there, though. Then, this question will become relevant again)

People will only resort to using bots if the game is a pointless grind and the price goes to the best grinder. You're not going to avoid bots on a game like that.

  • \$\begingroup\$ Can you go into more detail? What properties would a game need to have in order to be hard to write a bot for? \$\endgroup\$
    – Philipp
    Jan 2, 2019 at 13:07
  • \$\begingroup\$ A game of pure chance would be impossible to write a bot for. As would a game based on anything that AI is not yet capable of doing (well) e.g. natural language, or complex visual processing. "Click on the Red Duck hidden behind a blue wall" combines both, and would be something a 3 year old human could manage but you'd need some of the most advanced AI on the planet to even understand the question. (we're getting there though!) \$\endgroup\$
    – JeffUK
    Jan 16, 2019 at 13:02

It's not easy, and its feasibility depends on your technical ability and funds. It also depends on the game itself. e.g. if it's a game of pure chance, writing a bot won't help at all. But if it's a game of speed, then obviously a bot will trump any human.

It's perfectly possible though, there are hundreds of online casinos and poker sites, all of these host online games which give cash prizes, and plenty more online non-cash games which implement successful bot prevention strategies.

They have sophisticated anti-cheat technology tuned to the game in question to prevent both automatic and 'manual' cheating (e.g. Poker needs collaboration detection as well as basic web application security measures etc. etc.)

You can also integrate some form of CAPTCHA into your game play; detect human-like mouse movements and non-linear timings etc.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .