Most shopping centres have entrances near the front and are designed to make you circle out toward the far spaces. So everybody ends up prudent or optimistic (== those that circle back hoping someone left since they last drove past the entrance)
I'm the optimistic driver and the childish enjoyment of finding that spot near the front when the whole parking lot is full is worth the back tracking when it fails. In reality the meek approach isn't really terrible in the grand scheme and you get more exercise.
The naive answer is 1/e (where e is the base of the natural logarithm). If you have N applications, reject ~N/e applications and choose the first applicant who is better than every previous applicant.
It’s similar but not quite the same. In the parking problem you don’t know how many free spaces there are to begin with, and the ‘quality’ metric of each space is just a linear function of its location. In the idealized problem presented, each space as you go along the line is strictly better than every previous space - you just don’t know whether it is the last one.
The analogous secretary problem would be to be presented with a sequence of increasingly suitable candidates, but never knowing whether each candidate is the last one.
I had the same feeling. If memory serves, the problem of optimal stopping is discussed (using a parking lot as an example) in the book "Algorithms to live by"
I always try and park away from the busiest areas for the same reasons but, purely from observation, it's obvious most people don't since the area closest to the store entrance is almost always utilised the most.
I do often wonder why people are so lazy or so busy that saving on walking an extra few yards is so important to them?
When our baby was younger we tended to park as far away from anyone else as possible to give us room to put her back in the car. It's amazing how often we come back to find someone parked right next to us.
I've noticed this as well. If I park in the middle of the parking lot with no cars anywhere near me, very often I come back to find another car parked right next to mine.
My theory as to why this happens is that a parked car is a clearer indicator of how to park properly. If I try to park properly using the painted lines (which are often not very visible and go out of sight when you start entering your parking spot) I might miss the mark. However, a parked car that's more visible is a better indicator of whether I've parked correctly.
One of the reasons I back into parking spots. Back in, watch your mirrors, line up the lines using your mirrors or backup camera and continue going back. You will park perfectly in the middle every time.
For safety you should always park so you drive out of your parking spot. If you back out you risk that you don't see some kid walking right behind your car. Particularly if when you live a couple trucks/vans have parked next to you so there is no way to see if there is anything else coming and you have to guess.
Correct! And if we are listing reasons for backing up: Because you have to pass the parking spot before you back in, you can see if the spot is clear of debris, shopping carts and children.
> I always try and park away from the busiest areas for the same reasons but, purely from observation, it's obvious most people don't since the area closest to the store entrance is almost always utilised the most.
Walking through parking lot is not ideal. If it's outside, then you need to be dressed out for walking outside (and be prepared for rain etc. when coming back). When it's inside, it can smell of fumes and not neccesarily be very safe.
I’d like to guess that it isn’t a well-thought decision to park closer. It’s just basic instinct of choosing a convenient place. While in reality that causes crowdedness and might cost more time, the drivers aren’t doing a rational computation in their head about optimization. If you ask them the question, “Is that few steps so important to you?”, they will answer no and park a bit further.
I've never been called 'meek' before! But then I think I'm looking at the problem differently. I don't consider the walking time wasted, I don't mind the walk. I want to get the car parked as quickly as possible, not spend the least amount of time between entering the lot and getting to my destination. In fact, I will often park further away even when I can directly see that close up spots are free.
I would even consider the short walk to be a bonus! Also, you do not have to squeeze your car into a narrow slot, but can choose a spot with another free slot to your driver side -- or both sides, if you have a passenger!
1) no need to worry about someone pulling out in front of me or pedestrians or carts hitting my car
2) I’m at a desk all day. A couple minutes walk every time I park adds up to count as exercise
Ah, yes, well for most indoor parking garages my usual strategy is to immediately head for the top (or bottom) floor. Almost nobody parks there unless they have to, so there's often lots of open spaces and much less traffic.
If there aren't and good spots on the last level, then there likely aren't any good spaces anywhere in the garage anyway.
Curve-ball: the mall I visit most often has separate entrance and exits. The best spot equidistant from both. I've spent way too much time on this problem in my head.
In multi-level car parks I believe people overestimate vertical distance and underestimate horizontal.
In the carpark I park in, I often just speed down to the level where there first begins to be car spaces and I can often get much closer to the stairs which lead back up to the exit on the ground floor.
Extra vertical distance is only 4-5m but saved horizontal distance can be 50m.
This is true, the car park has a ramp system that is all in one corner of the car park but I guess most places make you drive to the other side to go up one
OK, I admit it. I use the meek strategy. Why? because I believe the parking spots that minimize walking from car to destination should be left empty for people who need them more than I do.
On the rare occasion where I need to minimize walking (because I have something heavy to carry, for example) I park to minimize walking even if I have to waste some time and fuel doing it. In that case it helps me when other people use my default strategy.
Optimization-theory work like this paper has a fundamental flaw: it assumes the actor is a classic "econ" -- the ideal person (defined by classical economics) working to optimize purely personal utility. In fact, many people work to optimize community utility.
Studying that community-optimization model is harder. But it is probably worthwhile. Such large-scale systems as autonomous vehicles will require community optimization to succeed.
It's time to reread Kahneman and Tversky, and Richard Thaler.
I actively want the exercise. I take the stairs at work rather than the elevator when practical for the same reason.
I don't think it's fair to categorise being a decent person as "meek". If more people did that, we'd have a far lower need for things like disabled and family spaces because people would fill inwards.
But the social-science and system-science issues get confused by using ethically charged words like "decent." Modeling "decent" gets confusing without ways of measuring it. (I'll spare us a long digression on how slippery it can be to understand ethical / moral reasoning.)
Do I leave parking spots for other people because I'm a "decent" person? I like to think so -- I like to think I put others' needs ahead of my own when I can -- but that's my business. From a systems point of view I do it to help control parking lot congestion.
If I were a business owner and didn't leave convenient parking spots for others, I'd be deliberately adding to congestion near my business. That's a cost. What's the benefit?
Example: Van Jacobson's slow-start / exponential-backoff algorithm to help prevent TCP endpoints from saturating routers. It's certainly possible to explain this to a roomful of undergrads as a "be nice" requirement on web servers and other sources of large TCP streams. But it's more productive to model it as part of a large-scale system to control congestion.
"However, if there a few meek drivers while the majority follow the prudent or optimistic strategy, then the meek strategy is not bad because meek drivers will park
a distance λ from the target."
I am a 'meek' driver for this very reason. My assumption is that the majority of other drivers are either 'optimistic' or 'prudent,' and as such the parking cost of the meek strategy is much lower. Would be very interesting to test this in the real world!
There are a few parking lots at my workplace, and I park at one fairly close, at a time where the last spots usually fill up.
I haven't kept a tally, but there have been multiple times (after a few earlier experiences on the other end of the spectrum) where I've gotten the last spot on the far end of the lot, only to see the few cars ahead of me drive around and eventually end up going off to another lot.
In my case I'm the meek driver on a micro/lot level, but a prudent/optimistic one on a macro/workplace level. But in my anecdata it's probably saved me more time than any other behavior I could employ.
I am wondering how did the "meek" strategy look like in detail.
Because, if you try to occupy the very first spot available, then of course it makes very little sense. But I believe that a "human meek" would instead skip the whole line of unoccupied spaces and take the last spot in such a line. This "unoccupied line" does not even have to begin with the first parking space, but for example with the third. Though you need to assume that the line of empty spots is in driver's line of sight.
This is great. Fortunately, at my work, if I just drive past the main entrance, and enter the parking lot nearest our entry door, I can start my journey through the lot at the closest point. This obviously eliminates the need for guessing or betting or settling early.
I admit that at certain times I do enter from the furthest entrance...and I always drive past the first spot, and bet on that second one. Quite often there is a better spot as I'm walking in, but sometimes there's not. So when faced with the opportunity, I follow the prudent strategy.
Now at an unfamiliar location (walmart, movie theater, whatever), I always try to find a spot that butts up against a curb (preferably on one side AND the back of the spot as well) to lessen exposure of my vehicle to other motorists (doors, bumpers) as much as possible.
This actually suggests that there are interesting psychological implications to parking lot circulation design. If you design a lot so that people can explore it in an order that matches a monotonous decrease in the ‘value function’ they place on each space location, you reduce driver stress because they can take the simple strategy of parking in the first available space and know they could not possibly have done better by following a different strategy (ah - except for waiting by the entrance and holding up traffic til someone leaves, of course).
This assumes you can identify a common value function everyone (or most people) places on each space of course.
Making people follow the value function in increasing order - as in the example, where each space you pass is closer to the entrance and so ‘better’ - seems to force this strategy choice.
This suggests all kinds of analogies in user experience and organizational and game design problems...
Hmm - do non-mathematicians "optimise" their parking? I really doubt it.
I feel like the people trying to park right near the door/entrance to the shops are not "optimising" the time they're there - I feel like they're probably just being lazy and "optimising" for the smallest amount of energy exertion possible (consciously or not). After all, why walk when you have a car? They don't seem to mind spending time following people walking back to their cars, or just looping around and around, or trying to squeeze into spaces that are too small when it is clearly obvious that you can just pull straight into a space another 20-30m away without any waiting or tight squeezes, then spend 15-30 seconds more walking. They're wasting a lot more time trying to get close so it can't be time they are optimising for otherwise people wouldn't do it.
(some people of course can't walk distance but then there are disabled spaces, parent + child spaces etc)
This model assumes hiigh availability of spaces - and that's not what I see very often outside the wide open spaces of America.
A better model would include time to find a free parking space. Searching and queuing time can quickly dominate under high utilization. Taking the first available space is very far from "meek" when you have to struggle to get it.
Of all the metrics in my life, optimizing for minimum car-to-store time is not important or even a consideration for me.
Unless it's pouring rain, I'll usually park near the back of the lot and take my time walking to the store. Even if it's raining, I won't usually opt for the spots closest to the store. I don't know _why_ that latter behavior exists, but it's what I do.
I park to optimize my exit from the parking lot. Pull-through spots to avoid reversing, avoiding needing to perform left turns, things like that. If I'm at the food store I'll also try to be near a cart return area, if possible.
I go by hypermiler advice, optimize distance traveled by car.
Choose spot closest to entry-exit path of car park. Park closer to exit, and so that you don't need to reverse out, minimize inefficient slow/idle movement with cold engine.
And any additional walking is a bonus as mentioned by several people here.
I'd like to see a version of this that optimises for this problem more globally. Other people exist!
e.g. Park further away if you can handle it (arrived on time, and are in good health, and have little luggage). Then people who are struggling (arrived late, with kids, in poor health, with lots of luggage) have a higher chance to find a space closer to the entrance. This might invisibly make someone have a less shitty day, and reduce my own risk of running into an angry person later. It's more impactful to optimise for that, rather than save seconds of walking.
52 comments
[ 2.7 ms ] story [ 111 ms ] threadThe naive answer is 1/e (where e is the base of the natural logarithm). If you have N applications, reject ~N/e applications and choose the first applicant who is better than every previous applicant.
The analogous secretary problem would be to be presented with a sequence of increasingly suitable candidates, but never knowing whether each candidate is the last one.
https://www.santafe.edu/news-center/news/where-how-park-your...
http://algorithmstoliveby.com/
I do often wonder why people are so lazy or so busy that saving on walking an extra few yards is so important to them?
My theory as to why this happens is that a parked car is a clearer indicator of how to park properly. If I try to park properly using the painted lines (which are often not very visible and go out of sight when you start entering your parking spot) I might miss the mark. However, a parked car that's more visible is a better indicator of whether I've parked correctly.
Walking through parking lot is not ideal. If it's outside, then you need to be dressed out for walking outside (and be prepared for rain etc. when coming back). When it's inside, it can smell of fumes and not neccesarily be very safe.
Being able to find your car when you leave.
Minimizing exit time when leaving an event where everyone gets out at the same time.
Being able to see the whole (or enough of the) lot upon entering-- so you don't have to guess.
1) no need to worry about someone pulling out in front of me or pedestrians or carts hitting my car 2) I’m at a desk all day. A couple minutes walk every time I park adds up to count as exercise
With that in mind, I usually opt to carry the grocery out myself because I prefer it. But rationally, the cost disappears because of the option.
In the carpark I park in, I often just speed down to the level where there first begins to be car spaces and I can often get much closer to the stairs which lead back up to the exit on the ground floor.
Extra vertical distance is only 4-5m but saved horizontal distance can be 50m.
Indeed, for the walking component. However having to drive up and then back down additional levels has time and hassle costs too.
On the rare occasion where I need to minimize walking (because I have something heavy to carry, for example) I park to minimize walking even if I have to waste some time and fuel doing it. In that case it helps me when other people use my default strategy.
Optimization-theory work like this paper has a fundamental flaw: it assumes the actor is a classic "econ" -- the ideal person (defined by classical economics) working to optimize purely personal utility. In fact, many people work to optimize community utility.
Studying that community-optimization model is harder. But it is probably worthwhile. Such large-scale systems as autonomous vehicles will require community optimization to succeed.
It's time to reread Kahneman and Tversky, and Richard Thaler.
I actively want the exercise. I take the stairs at work rather than the elevator when practical for the same reason.
I don't think it's fair to categorise being a decent person as "meek". If more people did that, we'd have a far lower need for things like disabled and family spaces because people would fill inwards.
But the social-science and system-science issues get confused by using ethically charged words like "decent." Modeling "decent" gets confusing without ways of measuring it. (I'll spare us a long digression on how slippery it can be to understand ethical / moral reasoning.)
Do I leave parking spots for other people because I'm a "decent" person? I like to think so -- I like to think I put others' needs ahead of my own when I can -- but that's my business. From a systems point of view I do it to help control parking lot congestion.
If I were a business owner and didn't leave convenient parking spots for others, I'd be deliberately adding to congestion near my business. That's a cost. What's the benefit?
Example: Van Jacobson's slow-start / exponential-backoff algorithm to help prevent TCP endpoints from saturating routers. It's certainly possible to explain this to a roomful of undergrads as a "be nice" requirement on web servers and other sources of large TCP streams. But it's more productive to model it as part of a large-scale system to control congestion.
"However, if there a few meek drivers while the majority follow the prudent or optimistic strategy, then the meek strategy is not bad because meek drivers will park a distance λ from the target."
I am a 'meek' driver for this very reason. My assumption is that the majority of other drivers are either 'optimistic' or 'prudent,' and as such the parking cost of the meek strategy is much lower. Would be very interesting to test this in the real world!
I haven't kept a tally, but there have been multiple times (after a few earlier experiences on the other end of the spectrum) where I've gotten the last spot on the far end of the lot, only to see the few cars ahead of me drive around and eventually end up going off to another lot.
In my case I'm the meek driver on a micro/lot level, but a prudent/optimistic one on a macro/workplace level. But in my anecdata it's probably saved me more time than any other behavior I could employ.
Because, if you try to occupy the very first spot available, then of course it makes very little sense. But I believe that a "human meek" would instead skip the whole line of unoccupied spaces and take the last spot in such a line. This "unoccupied line" does not even have to begin with the first parking space, but for example with the third. Though you need to assume that the line of empty spots is in driver's line of sight.
I admit that at certain times I do enter from the furthest entrance...and I always drive past the first spot, and bet on that second one. Quite often there is a better spot as I'm walking in, but sometimes there's not. So when faced with the opportunity, I follow the prudent strategy.
Now at an unfamiliar location (walmart, movie theater, whatever), I always try to find a spot that butts up against a curb (preferably on one side AND the back of the spot as well) to lessen exposure of my vehicle to other motorists (doors, bumpers) as much as possible.
This assumes you can identify a common value function everyone (or most people) places on each space of course.
Making people follow the value function in increasing order - as in the example, where each space you pass is closer to the entrance and so ‘better’ - seems to force this strategy choice.
This suggests all kinds of analogies in user experience and organizational and game design problems...
I feel like the people trying to park right near the door/entrance to the shops are not "optimising" the time they're there - I feel like they're probably just being lazy and "optimising" for the smallest amount of energy exertion possible (consciously or not). After all, why walk when you have a car? They don't seem to mind spending time following people walking back to their cars, or just looping around and around, or trying to squeeze into spaces that are too small when it is clearly obvious that you can just pull straight into a space another 20-30m away without any waiting or tight squeezes, then spend 15-30 seconds more walking. They're wasting a lot more time trying to get close so it can't be time they are optimising for otherwise people wouldn't do it.
(some people of course can't walk distance but then there are disabled spaces, parent + child spaces etc)
Unless it's pouring rain, I'll usually park near the back of the lot and take my time walking to the store. Even if it's raining, I won't usually opt for the spots closest to the store. I don't know _why_ that latter behavior exists, but it's what I do.
Choose spot closest to entry-exit path of car park. Park closer to exit, and so that you don't need to reverse out, minimize inefficient slow/idle movement with cold engine.
And any additional walking is a bonus as mentioned by several people here.
e.g. Park further away if you can handle it (arrived on time, and are in good health, and have little luggage). Then people who are struggling (arrived late, with kids, in poor health, with lots of luggage) have a higher chance to find a space closer to the entrance. This might invisibly make someone have a less shitty day, and reduce my own risk of running into an angry person later. It's more impactful to optimise for that, rather than save seconds of walking.
Any ideas for how to math this?