I’ve genuinely been enjoying this game, finding it quite engaging and even insightful in many ways. While some levels, in my opinion, might unintentionally teach some questionable coding practices (perhaps a discussion for another time), overall, the game mechanics have felt well-implemented and rewarding. However, I’ve hit a significant roadblock with level RL-1, and it’s causing considerable frustration. Level RL-0, while initially a bit puzzling, became clear after carefully rereading the instructions, allowing me to successfully navigate it without major issues. RL-1, on the other hand, provides virtually no clear guidance, leaving me completely bewildered.
Firstly, the lidar sensor placement on the vehicle in RL-1 appears fundamentally flawed. It’s depicted as a small box situated directly on the car’s bumper, yet it seems to register objects as if they are several meters ahead. This discrepancy means the vehicle inevitably collides with obstacles before the lidar can even gather meaningful environmental data. I strongly suspect this is a bug, and the sensor is intended to be positioned much further forward, similar to its placement in RL-0. You can see an example of this issue in this screenshot:
I’ve come across older forum discussions mentioning ‘adjusting lidar placement’ by ‘clicking’, but despite thoroughly clicking every element of the interface, including repositioning the sifter module in case it was related, the lidar configuration remains unchanged. Am I missing something? Is lidar adjustment supposed to be a feature? The game hasn’t hinted at this possibility at all, and the interface doesn’t seem to offer any such controls. This lack of clarity is hindering the “True Learn” experience the game should be offering.
Then there’s the machine learning aspect, which, frankly, appears to be completely non-functional in this level. Regardless of the initial population size or selection percentage I set, the three epochs consistently play out identically on my screen. This consistency persists even when I adjust these parameters between epochs. Each epoch concludes with the message ‘you hit the target speed!’, which seems detached from the actual outcome. When I initiate a test run, the vehicle performs the exact same action every single time: it abruptly slams on the brakes and comes to a complete stop. This behavior is so predictable, even across 50 attempts with occasional random seed adjustments, that the final state is always identical. What’s even more puzzling is that accident avoidance doesn’t seem to be an explicitly stated objective; the mission description suggests the car simply needs to achieve the highest possible speed. This makes the consistent braking behavior even more illogical and counterintuitive to what one would expect from a “true learn” simulation.
The mission, as it stands, feels impossible to complete, leading me to believe it’s fundamentally bugged. Compounding the issue is the mission’s ambiguity regarding the tools and their intended effects. This vagueness has forced me to invest a significant amount of time experimenting with various approaches, attempting to convince myself that I’m not simply misunderstanding the mechanics. Unfortunately, this prolonged troubleshooting has significantly diminished my enjoyment of the game. The intended “true learn” moments are overshadowed by frustration and confusion.
Could someone confirm whether RL-1 is indeed broken? Thankfully, it doesn’t seem to block game progression, but it’s currently preventing me from recommending the game to others. The current state of RL-1 doesn’t feel like a valuable learning experience, and resolving this issue would greatly enhance the game’s appeal and its ability to deliver on its promise of “true learn” through gameplay.
