4 ms·

You would only find that if you were literally flooring it to merge. That's not safe regardless.

tacitusarc 4y ago
Have you really never been in a situation where you need to accelerate out of the way of another driver braking the law? For example, driving through an intersection and suddenly you notice someone is running their red light on a collision course. You must either go forward or backward. Much more difficult to reverse than to accelerate.
- googlryas 4y ago
  I check intersections before proceeding through them , so no that hasn't happened in 250,000 miles of driving. And what is that, like 0.5 seconds to make a difference? How much further will the car be when it had this subscription vs not having the subscription?
  - hansvm 4y ago
    Half a second seems roughly right. An exact answer depends on the car's power band and other more complicated factors. You'd expect an extra 2-8 inches, and if the interesection were a bit wider and slower so that you had 1.0 seconds you'd expect an extra 16-64 inches. Also, on such small timescales the engine's responsiveness would come into play, and it's unclear if the tuned package would affect that for sure, but it'd be hard for it not to. 2-8 inches probably won't give you better odds of any meaningful outcomes beyond an extra few percent. 16-64 inches could easily be the difference between a miss and a crash (especially if the red-light runner swerves or counting the selection bias in which distribution of possible strike zones you're likely to have predicated on the assumption that accelerating was a reasonable course of action), or would very likely be the difference between an extremely dangerous collision with the driver's door and something more reasonable toward the back of the car (worse with kids in the back -- thankfully for the purpose of this discussion, only 15% of cars in the US have passengers).
hansvm 4y ago
That depends how the UI to the extra power is implemented. I bet the lower power mode still has increasing power the further the pedal is pressed (as opposed to simply finding there is no extra power if you floor it -- a smooth monotonic graph rather than one clipped at the top), so you still have a feedback loop going through the latency of conscious thought as you press the pedal the "right" amount to accelerate (not flooring it), find that it's insufficient, and now still have to choose between braking and accelerating -- either braking after you've already decided that's the less safe option and accelerated enough that that's probably more true than it was previously, or accelerating now that you've wasted some time and need to accelerate more briskly than you previously would have (maybe even flooring it, which as you say is not a safe place to be) while also having doubts about whether attempting to accelerate faster will actually suffice since your current pedal input should have also sufficed. Edit: Not to mention, lack of safety in an absolute sense doesn't mean that the safest option for a given driver (predicated on the assumption that it looked safe to go as far as the onramp and there's no way off other than merging eventually) wouldn't be to floor it to merge. In the Bay Area in particular, drivers are notoriously aggressive about both tailgating and blocking merge attempts, and faced with a tricky merge you might find that braking input is unlikely to work (no openings toward the rear), heavy braking to avoid the situation is unlikely to work (in theory you're supposed to yield to the freeway traffic, but doing so will probably result in the tailgater behind you rear-ending you), and the nearest available opening in front of you requires heavy acceleration. It'd be better if everyone could be polite and zipper merge, but in a bad situation with limited personal control you can only hope to make the best of it.

Have you really never been in a situation where you need to accelerate out of the way of another driver braking the law? For example, driving through an intersection and suddenly you notice someone is running their red light on a collision course. You must either go forward or backward. Much more difficult to reverse than to accelerate.

googlryas 4y ago
I check intersections before proceeding through them , so no that hasn't happened in 250,000 miles of driving. And what is that, like 0.5 seconds to make a difference? How much further will the car be when it had this subscription vs not having the subscription?
- hansvm 4y ago
  Half a second seems roughly right. An exact answer depends on the car's power band and other more complicated factors. You'd expect an extra 2-8 inches, and if the interesection were a bit wider and slower so that you had 1.0 seconds you'd expect an extra 16-64 inches. Also, on such small timescales the engine's responsiveness would come into play, and it's unclear if the tuned package would affect that for sure, but it'd be hard for it not to. 2-8 inches probably won't give you better odds of any meaningful outcomes beyond an extra few percent. 16-64 inches could easily be the difference between a miss and a crash (especially if the red-light runner swerves or counting the selection bias in which distribution of possible strike zones you're likely to have predicated on the assumption that accelerating was a reasonable course of action), or would very likely be the difference between an extremely dangerous collision with the driver's door and something more reasonable toward the back of the car (worse with kids in the back -- thankfully for the purpose of this discussion, only 15% of cars in the US have passengers).

hansvm 4y ago

That depends how the UI to the extra power is implemented. I bet the lower power mode still has increasing power the further the pedal is pressed (as opposed to simply finding there is no extra power if you floor it -- a smooth monotonic graph rather than one clipped at the top), so you still have a feedback loop going through the latency of conscious thought as you press the pedal the "right" amount to accelerate (not flooring it), find that it's insufficient, and now still have to choose between braking and accelerating -- either braking after you've already decided that's the less safe option and accelerated enough that that's probably more true than it was previously, or accelerating now that you've wasted some time and need to accelerate more briskly than you previously would have (maybe even flooring it, which as you say is not a safe place to be) while also having doubts about whether attempting to accelerate faster will actually suffice since your current pedal input should have also sufficed. Edit: Not to mention, lack of safety in an absolute sense doesn't mean that the safest option for a given driver (predicated on the assumption that it looked safe to go as far as the onramp and there's no way off other than merging eventually) wouldn't be to floor it to merge. In the Bay Area in particular, drivers are notoriously aggressive about both tailgating and blocking merge attempts, and faced with a tricky merge you might find that braking input is unlikely to work (no openings toward the rear), heavy braking to avoid the situation is unlikely to work (in theory you're supposed to yield to the freeway traffic, but doing so will probably result in the tailgater behind you rear-ending you), and the nearest available opening in front of you requires heavy acceleration. It'd be better if everyone could be polite and zipper merge, but in a bad situation with limited personal control you can only hope to make the best of it.