Idle Machine Learning
Update - October 30, 2019:
You can download the game and use it on your own web as long as you do the following:
- Do not modify the game.
- Playing the game must be free (that includes not needing a paying subscriptions to play).
- Using your own ads in your web is allowed.
- Give credits (attribution) to Vashgard.
You can contact me at vashgard@gmail.com
Let your computer learn to drive racing bikes!!
How does Machine learning works?
A wave of motorcycles will be created, each one making RANDOM decisions based on its sensors.
One of them will be the one that has made the best decisions and the one that has gone the furthest before crashing.
You will have to wait 3 seconds between waves (You can reduce it by using "gen. Time" skill).
The next wave, will imitate the decisions of the best motorcycle of the previous wave, but with small RANDOM modifications.
One of these small RANDOM modifications can cause one of the bikes to go a little further before crashing.
When this happens, the whole group learns for the next wave, and so on until they learn to complete the circuit.
Good luck!
Download
Click download now to get access to the following files:
Development log
- Medals!Nov 13, 2019
- Available to download and useOct 30, 2019
Comments
Log in with itch.io to leave a comment.
Just beat level 9! Really needs an ending screen of some kind.
well, the drivers are definetely artificial.
I kinda wish the game leaned more into the AI aspect, like changing the learning rate etc. a button to reset current inteligence without retiring would also be nice(I know that the game is old, but that one would be really appreciated)
It is so terribly annoying when they just spend 50 generations bashing their heads against the exact same rock.
You have plenty of steering, SO STEER. Stop speeding at the rock, YOU HAVE A BRAKE SO USE IT.
Why are you actively steering INTO THE WRONG DIRECTION after 30 generations of suicide? I lowered your top speed so you'd have a better chance of NOT HITTING THE EXACT SAME ROCK AS YOUR GREAT*30 GRANDFATHER.
I know the game CLAIMS they learn better after retiring once and getting a 2nd sensor, but these lemmings only seem to get dumber. Maybe that 2nd sensor is more curse than blessing?
hahahaha I made this game many years ago, and I still enjoy getting comments on it! ^^
Maybe it was badly explained on the game (I don't remember now) but second and third sensor don't make it learn faster, otherwise it makes they learn slower, but able to learn more and better (in machine learning terms, they have bigger layers).
Sometimes learning process gets stuck on a "local minimum" and it is very hard to get out of that. The only advise I can give you is to made they to "unlearn" a lot to get away of that "local minimum", and then start again the learning process. To do that, get back to old maps for some generations, or change the configurations by A LOT.
The green bike is there to allow you to see if each generation is evolving or they are stuck on a local minimum and getting away from it: if the green bike is always doing the same, they are not learning, neither "unlearning".
Another hard option you have is to retire early, you won't get the bonuses from completing all races, but you will start again with the same amount of sensor to try again learning from 0...
it learned to go backward?!?!?!
please show cost for points and other stuff you buy
never mind its at the bottom
The only things I'd add are:
1: After completing level 9 stats uncap to 10
2: No cap on money multipliers (you'd need to beat level 9 each time)
Why is something so simple so fun and entertaining?
a very good game!
I love how you need to think what will benefit your build most and the play with the sensors.
I think there can be an improvment with the speed as I at least didn't understand why you need to spend points in it but as this game is abandoned I guess it will never happen.
thanks for the wonderful game!
i beat it
Frustrating (and probably abandoned) game.I'm feeling insulted.
After having mastered it a few times (including an under 60 turns winning game
using the same or similar parameters on the same levels, I've never been able to replicate or even approach it - even after a 500 turns test) I'm forced to conclude that winning or losing mostly doesn't depend on my efforts but on the game's poorly managed randomness - and if we take into account the unobtainable ('coz bugged) medals, it's just...
... sad, it looked so funny & promising. I've read the dev writing about the tech he used (Unity) that 'this game is 4 years old xD'. Well we're a large bunch who are still playing 20 to 40 years old games built on clumsy and obsolete tech.
4/5 for the promising concept, .5/5 for the deceptive result. In the end, .5/5.
PS: about mastering a game (which can't happen here anyway since major bugs have never been fixed) if there's too much randomness there's no mastering at all and the player just owns a gilding arsehole.
Hi, I sorry you felt insulted.
I made this game "for fun" 5 years ago, and you are right, it is abandoned and it have bugs and things that can be improved, but I am not a game dev and I have other job. I just keep the game here because it makes me happy to find nice comments time to time from people that discovered the game and enjoyed it.
I understand and agree with your feed back, but I would like to say: try to be more constructive with the next feedbacks you gave to other real game dev, specially the ones that are starting.
A good constructive feedback would be something like "I think that if you change this, the game would be less random" instead of "this game is too random".
In this case, it is ok to me since I am not working on game dev. But reading harsh comments instead of constructive ones, can make junior Devs to give up instead to improve their skills.
Could you make this open sourced please! I'd love to mess around with it and learn more about ml
Sheesh you don't have to insult the developer so much man..
Mastered it
In this game, there is a hidden, uncontrollable variable named "Luck"
Love this idea, thanks for a fun game! Beat the last level, woot.
Does acceleration work? They only ever go minimum speed it looks like.
Ever thought about creating another idle game with machine learning like this? I'd love to play it.
Acceleration does work, unlike the bikers' 'decisions'.
I will build 10000 shop and a will take 1000000000000000000000000000000000000000000 hours
The medal is broken and doesn't activate.
if you want to see if you've beaten it check to see if you have a Pb
All of my cars just turn straight into the wall after the u-turn on the right in the third map. Is this a bug or should I just wait and let them learn? They've been doing it for a while now
This game is perfect while listening to Stayin' Alive.
careful with the unity bs
Don't worry, this game is 4 years old xD
I've completed all the levels, but didn't get the medal for it. It was very fun tho.
How do I save my game? It keeps rolling back progress every time I log on.
How do I fix my bots crashing on the 4th turn every time on level 4? They've been doing this for weeks
Try giving they a wider angle of vision, so they can see the open turn better:
Still not working. They crash right here:
Should I retire?
alright thats it im going back a level
To try to change a deep learned behavior, is good to go back a level and change a little the skills configuration. This will make they to fail the level in the first turns, but the longer you keep relearning the previous track, more they will lose of this bad behavior and get a better change to learn to do that turn.
If this didn't work, retire is a good option too, you won't need a lot of time to get back to lvl 4 again after retirement.
after all these yeas i have re found this fucking gam
:)
What is the green motorcycle? At first i thought it was a "freebie" unit that would never crash, but then it does crash sometimes. Is it the first motorcycle to succeed at completing the course?
I think the green bike is the one you can focus on from the start
It is the best one of the previous generation.
As reference, when the green one is better than the other bikes, your IA is not getting smarter. When other bikes does better than the green one, they are learning.
It would be nice if we could see the neural network, I can't manage to get the AI to learn to speed up on it's own.
LOL LVL 6 AT 19 GENS
Imma post this on github(to be hosted there, maybe its better than itch.io hosting) I'll post the link once I do it
Great game! One problem is the AI usually gets stuck doing one thing over and over, so a mutation upgrade could be cool. Otherwise, Great job!
The first retire was very easy, but the second, they got really stupid level 2 so I reset, and then the second retire they were practically omniscient and first tried all of them after the 2nd
Edit: And now I'm completely stuck at the 180 in level 9
I would just like to put it on the record that my bikes managed sub-9 seconds on the first level :O Great game fella
for three laps then right
How long does it usually take to finish the game?
Great game, went back to level one after 300+ generations to get 11.49
I just found this gem and it's so simple yet so funny. Can't stop playing. Thank you for sharing this game :)
I completed level 9 but I did not get the medal. Im not mad, just disappointed.
Me too! I want to be acknowledged!
yeah right
I came to say the same thing, I had a bike finish 3 laps. maybe you need more than that for it to count?
With medal? the "It is posible!" medal?
You have to retire for third and last time after finishing lvl 9 to earn it.
Very interesting concept.
I would like to know how the specific evolution algorithm works because, observing those bikes' behaviour, it doesn't look like there's an actual neural network behind but something more like a sequence of actions they learn to do.
HOLY- I JUST STARTED AND I GOT 3 LAPS IN 2 TRIES. THESE THINGS ARE ALREADY SENTIENT, NO LEARNING REQUIRED
I regret saying that, I've had to retire multiple times before level 3 because level 2 bikers were stupid lmfao
i completed level 9 in 2:17.441...w/o medal. could be a bug?
I completed in 2:18 and didnt get a medal neither
how to get every achievement
i did it: play the game
they are learning: play the game
piece of cake: play the game
my first retirement: have nothing better to do
one last time: really have nothing better to do
it's possible: really really have nothing better to do
avenge my son: have max speed
my great great great grandfather: have max speed
we are legion: have max speed and generation time
these bikes are dumb: have high population size
this is not machine learning: have max population size and high speed
omfg this game is so broke: have max population size, max speed, and high generation time
so 8 is the limit: play the game
almost there: play the game
money is useless now: play the game
the first one is free: why do you need a tutorial for this
tweet tweet: pres the twitter button
facebook ok boomer: press the facebook button
it literally tells you this
(After 1st retire)
Map 2 1st time: All but one dies and the last one dies before the finish.
Map 2 2nd time: Only 3 die and the rest make it all 3 laps.
How do you do the hairpin turn in map 9?
I have the same issue.
no idea how, but I managed to train the bikes
…to go backwards🤷♂️
i trained them to commit suicide after the 2nd turn
show me