mvnForum Homepage


Print at Aug 24, 2019 11:33:58 PM

Posted by tanonev at Mar 27, 2012 12:33:58 PM
Re: Botting
 
I do also have to point out that you don't have to understand the game too terribly well to write the bot, because if you have to skill to write one, you probably also have the skill to decompile the game and look at how the scoring works instead of guessing based on experience.


I'd like to point out that knowing the rules for the game, while certainly useful, is a relatively insignificant part of understanding the game well enough to write a bot for it. Take Go as an example; all of the rules are publicly known, but that doesn't mean that anyone currently understands Go well enough to write a top-level bot for it.

In fact, you can bypass knowing the scoring for the game just by using what other players say about the game. If other players say that making 3x3s is enough for a yellow pump in bilge, then all you need to do is figure out how to make 3x3s as efficiently as possible (either yourself or with a bot), and then follow that blindly. Similarly, if someone posts a sufficiently detailed description of how to construct Sea Donkeys and Vegases, all you need to do is follow that description (either yourself or with a bot), and then follow that blindly.

 
Suddenly when writing a bot means $500 a month, you get a whole lot more motivated to do it. I don't know what the solution is, but I can see the problem getting a lot worse in the near future.


The solution is something that would be nice to have anyway but apparently isn't feasible with the amount of dev time currently invested in this project: More puzzles, more puzzle variations. The core puzzles haven't changed in years, and aren't expected to change for years to come. This means writing a bot is a one-time investment; once you've successfully written it, you can continue to profit off of it so long as it's still effective (and note that unless using the bot results in immediate banning, you'll still have potential clients willing to pay for the bot for a few days' worth of use). However, bots also tend to be very specifically tailored for a puzzle, and will NOT adapt to a change in puzzle mechanics without effort from the bot-writer. If the mechanics change (subtly) often enough, you not only keep the experience fresh for actual players, you frustrate the bot-writer by forcing him to constantly update his bot. Imagine the extreme case of releasing a brand new set of duty puzzles every month; bot writing would pretty much go out the window because by the time someone released a bot that performed passably well, the puzzle would be on its way out. Obviously, releasing a new set of puzzles every month is infeasible without massively increasing developer resources, but more modest investments in varying the puzzles will still hamper the bots (not to mention give players who like learning and adapting to puzzles more enjoyment out of the game).
----------------------------------------
Tanonev on all oceans; currently exploring Meridian.
Puppetar by Tilinka

Puzzle Pirates™ © 2001-2016 Grey Havens, LLC All Rights Reserved.   Terms · Privacy · Affiliates