Table II: Performance comparison of models regarding score and other metrics. The average is calculated per metric and shows the stability of the training for each of the T*D* settings. The values shown in the parenthesis for Score and Dist are the standard deviations of the models. Offroad, Otherlane, and Either are calculated with a threshold of 20% for each time-step.

Model	Type	Off-Road ↓	Other-Lane ↓	Either ↓	Success ↑	No Collision ↑	Score ↑	Dist [m] ↑
TGDG (ours)	Average	0.0%	11.1%	11.1%	96.8%	84.1%	0.90 (±0.07)	11132 (±703)
	Best	0.0%	2.4%	2.4%	100.0%	100.0%	0.99	11168
TGDE (ours)	Average	0.1%	11.6%	11.7%	89.0%	80.8%	0.86 (±0.073)	11392 (±232)
	Best	0.0%	2.5%	2.5%	100.0%	100.0%	0.99	11119
TEDE (ours)	Average	11.4%	29.8%	37.9%	34.5%	73.4%	0.57 (±0.3)	10209 (±1950)
	Best	0.0%	17.5%	17.5%	100.0%	100.0%	0.94	11227
TEDG (ours)	Average	11.5%	25.8%	36.3%	33.3%	78.3%	0.58 (±0.3)	11437 (±2201)
	Best	0.0%	22.7%	22.7%	100.0%	100.0%	0.92	11238
RL Dosovitskiy et. al (2017)		42.4%	21.0%	52.4%	41.7%	50.0%	0.46	8787
IL Codevilla et. al (2018)		8.3%	0.7%	8.4%	86.9%	90.5%	0.90	11293