The Fast Lane EV YouTube channel recently put Tesla’s Full Self-Driving (Supervised) system to a unique challenge: navigating Colorado’s notoriously steep and unpaved Lick Skillet Road. With no lane markings and a mix of cars, trucks, and cyclists, the test pushed the limits of Tesla’s camera-based driver assistance technology.
Testing Tesla FSD on Lick Skillet Road
The Fast Lane EV team selected Lick Skillet Road in Colorado for their experiment, a dirt road stretching about 1.6 km (~1 mile) with an average grade of 14.2% and a peak grade of 18%. Known for its challenging terrain and lack of pavement, the road is regularly used by vehicles and cyclists, making it a demanding environment for any autonomous system.
During the test, the Tesla Model 3 equipped with Full Self-Driving (Supervised) managed to stay on the correct side of the road throughout the climb, despite the absence of lane lines. The TFLEV crew followed in a Toyota Land Cruiser as a recovery vehicle but ensured the Tesla had no lead car to follow, relying solely on its sensors and software.
Handling Obstacles and Real-World Challenges
The Model 3’s FSD system demonstrated an ability to interpret the road’s width and boundaries, aided by visual cues like shoulders, guardrails, and surrounding trees. However, when the Tesla encountered cyclists ahead, it maintained a safe following distance but did not attempt to pass them automatically. The TFLEV team eventually intervened manually to overtake.
A notable limitation emerged when the system failed to recognize a pedestrian walking near the cyclists until the last moment. While no incident occurred, this highlighted the need for human oversight in complex, mixed-traffic scenarios. Despite these challenges, the Model 3 handled oncoming vehicles around blind corners without issue, consistently keeping to the correct side of the road.
Competitor Comparison
Tesla’s Full Self-Driving (Supervised) system is often compared to advanced driver assistance offerings from other automakers, such as Mercedes-Benz Drive Pilot, GM Super Cruise, and Ford BlueCruise. While these systems have demonstrated strong performance on highways and well-marked roads, few have been tested in off-pavement or unmarked conditions like those faced on Lick Skillet Road. This test highlights Tesla’s focus on vision-based autonomy, but also underscores the importance of continued development for complex real-world environments.
Historical Context
Tesla’s global sales have grown exponentially since 2013, reaching 1.79 million vehicles sold in 2024. However, Model Y sales saw a -3.1% year-over-year change in 2024, reflecting a maturing market and increased competition. The company’s ongoing investment in autonomous technology remains a key differentiator as the industry evolves.
Why this matters
This real-world test demonstrates both the strengths and current limitations of Tesla’s Full Self-Driving (Supervised) system. While the technology can handle challenging, unmarked roads under supervision, it still requires human intervention in complex situations involving pedestrians and cyclists.
As autonomous driving systems continue to develop, such tests provide valuable insights into their readiness for broader deployment, especially in regions with diverse road conditions.
What this means for buyers
For prospective Tesla owners, this test shows that Full Self-Driving (Supervised) can handle more than just highways and city streets, but it is not yet a substitute for attentive driving in all scenarios. Buyers should be aware that manual intervention may still be necessary, particularly on unpaved or poorly marked roads.
Those considering advanced driver assistance features should compare Tesla’s system with offerings from other automakers and consider their typical driving environments before making a decision.
Key Specs
- Test Vehicle: Tesla Model 3 with Full Self-Driving (Supervised)
- Test Location: Lick Skillet Road, Colorado
- Road Length: 1.6 km (~1 mile)
- Average Grade: 14.2%
- Peak Grade: 18%
- Recovery Vehicle: Toyota Land Cruiser
Source
Images sourced from: arenaev.com, unsplash.com, consumerreports.org
