• Welcome to SC4 Devotion Forum Archives.

new traffic experiments

Started by ldog, October 23, 2009, 06:16:28 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

ldog

#140
A little tidbit about speed, it has been stated in other posts that the speeds are in KPH. Then they are also about 10/16 of what I would generally consider reasonable realworld speeds (in other words multiply by 1.6)

1 KMpH = 1000Meters/60minutes= 16.666666666666666666666666666667 METERS PER MINUTE!!!
How big are our tiles in SC4? About that big.

Like many Americans, I have trouble relating the Metric system to the physical world. So I tend to think in SAE units when I need to eyeball something.
3.5 MpH sounds like a reasonable walking speed, however that is 5.6 KMpH.
5.6/3.5=1.6
If we take all the other speeds in the stock game and multiply them by 1.6, we get:
Steet 34 , road 50 , ave 64, highway 131, rail 176, subway 240, monorail 320 KMpH
or
Steet 21, road 31, ave 40, highway 82, rail 110 , subway 150, monorail 200 in MpH

()what() ???  :o
??? So the speeds Maxis tried to use are actually in MpH.
???
:-\

Once again, someone else may have said this. I remember reading a post here and it was concluded the speeds were in KMpH. I don't remember seeing anyone making that relationship.
At any rate, like I said in Gridlockcity, there's nothing wrong with double checking someone elses work.
Actually I had to do a doubletake, I thought maybe I just got confused going back and forth between metric and SAE, but the math checks out.
Not like it really matters anyway, I just thought it was interesting.

The important thing to remember is, it is tiles per time unit. km/h

So why that little exercise anyway? You seem to recall me saying take reality and throw it out the window.
I probably did say something like that. Still, even when we want to depart from reality, it is good to have a reference.
Actually though, I am looking to scale reality into SC4's reality (as long as we don't conflict the prime directive; the game must be fun).
When we want to scale, we need to have proper ratios. So I started putting a bit more thought into speeds.
Subway is way too fast. 60 MpH is more like it, and that is under optimal conditions. The NYC subway averages about 18 MpH (because of stops) or so I have read (I haven't been on the wretched thing in even longer than I've been on a bus) of course because traffic is so horrid, the subway is still the way to go.
Another strange irony of my life, the thing that is one of the banes of my existence is the thing I am trying so hard to make a more accurate representation of here. It makes me sick when I think about how many hours of my life have been wasted sitting in bumper to bumper traffic. I must need my head examined taking this on as a fun project.

Street and road are about right. Street speed limits tend to be about 15-25 mph, road 25-35. Ave...that's a bit on the low end. 40-55 is more like it. Highway, well highway speed limits here run the gammut from 55-80 (at least last I remember they used to be as high as 80 in the western part of the country, I've been back on the east coast for a good 9 years now)

Trains, I really don't have much experience with trains. A little googling and light reading, I think 110 is fair for the class of rail our rail is. Monorail, if we consider it HSR then 200 is fair. (Europeans would beg to differ I'm sure) I know there are much higher record speeds and indeed some of this information is a bit dated, but then in 2003 it was probably closer to the truth.

I haven't mentioned the El since I consider it an extension of the subway, of course one does not have to look at it that way or keep the speeds the same (I think Jason didn't, although probably because the light rail uses the El settings)
I have of course left bus out of the discussion since as we all know the bus speed being higher than car was part of a balancing act on Maxis part. Since you have to walk to get to the bus and walking is much slower, the bus needed some juicing up. The higher your max commute, the less advantage the bus needs and that is probably why others have slowed the bus down.
I've also not mentioned trucks and freight trains since their relationship is only to each other. However trucks go as fast as cars and the speed of 150mph for a freight train also is consistent with the same source that gave me 110 for a passenger train.

Of course I have gone into a bunch of useless theoretical babble as usual.
It doesn't matter what you set your speeds at, you won't see it in game, except the commute time graph, which we can rescale to whatever we want, and of course it goes crazy when we have commuters anyway.

The useful thing here is that aside from the subway I find the ratios to be good enough for a starting point. (y'all like that bold? I think I'm going to start doing that around when I make a point so that people who don't want to read through all my pages of droning on can skip to the point) It was something Maxis didn't bone up too badly.
We will probably tinker with these ratios later in order to keep balance among transit types as we like. We may even wind up with the same speeds as A or B or Z or something completely different.
If we give the subway a much higher network capacity, then we can set the speed to 60 and still keep it competitive (with badly congested roads, but then if your road network is not badly congested, why are you building subways anyway?) Hell, maybe, just maybe we could set the subway to not have congestion, it certainly would make the congestion map a lot more readable.

ldog

#141
So tonight I decided I wanted to do some testing to find out the exact effects of the trip starting time, and to further see if I can get the stations to work properly with an entry cost of .96/walking speed.
I took set all travel preferences to 100% fastest method. Took out the CvS and intersection/turn effects as well (set everything to 1.0 basicly neutralizing those vars).
Interesting effect. Of course with all roads (I did this in a medium tile on a new city) and these settings effectively the pathfinder does no work, because the shortest route is also the fastest route.
A couple side effects though, I got a commute time of 0. The congestion display always shows green. The "chaos of cars" and the like messages will showup when the volume is at 600% (like on my busiest intersections). The worst intersections had combined morning/commute volume of over 9000 cars, this is more than 900% capacity. While I have not tested this, I have a suspicion that there is a ceiling on volume. (the same way as transit switches will max out eventually at some point)The game also ran extremely fast on Cheetah speed, I couldn't even tell when the traffic simulator ran by speed slowdown as you normally can.

Now for the second (very short) test I changed only the CvS to go to .3 speed at 300% congestion. On the first run all congestion everywhere went to red. By the second run of the simulator I started getting massive abandonment. Average commute times also instantly shot up.

Third test I am put a more normal CvS curve back in, 0%/1.3, 100%/1.0, 200%/.65, 300%/.30. Which of course put the city back to normal. Average commute quickly leveled out at 2.75 (I believe I have it scaled to 1:1) and I let the city run a good while. It is now 46 years old and it has been pretty stable for the last 10 years but work comes early and I am tired so that will be it for the night.

Hopefully tomorrow I will be able to make some progress.

RippleJet

Quote from: ldog on November 15, 2009, 06:15:18 PM
1 KMpH = 1000Meters/60Seconds= 16.666666666666666666666666666667

1 h = 60 min = 3600 s ::)

1 km/h = 1000 m / 3600 s = 0.27777778 m/s

catty

Quote from: RippleJet on November 17, 2009, 11:28:40 PM
1 h = 60 min = 3600 s ::)

1 km/h = 1000 m / 3600 s = 0.27777778 m/s

And 0.27777778 m/s ties in with what the Prima guide has down as the walking speed across a tile



The Prima guide is worth getting if you don't already have it and its only $9.99 USD from Prima Games.

I'm starting to sound like a saleswoman, so I'm heading off to bed

Goodnight

:sleeping:
I meant," said Ipslore bitterly, "what is there in this world that truly makes living worthwhile?" DEATH thought about it. "CATS," he said eventually, "CATS ARE NICE.

z

#144
Thanks for yet another relevant citation, Cathy.  You're absolutely right, although it's a little confusing because the two .27's refer to two different things.  ???

I recently sent a PM to Lenny enumerating many and various sources that speeds are in kph and time is in minutes; this is one more source that I had forgotten about.  And I'll also second Cathy's recommendation about getting the Prima Guide.

ldog

#145
Quote from: RippleJet on November 17, 2009, 11:28:40 PM
1 h = 60 min = 3600 s ::)

1 km/h = 1000 m / 3600 s = 0.27777778 m/s

:angrymore:

Uhhhhhhh...big oops

:-[

But keep reading, because my embarassment is over a typo I made in sharing the information, not an error in my calculations. I also went back and made the needed corrections, so what I stated is correct and consistent with itself.

catty

Quote from: z on November 18, 2009, 03:27:16 AM
...You're absolutely right, although it's a little confusing because the two .27's refer to two different things...

I know, It was way past my bedtime when I posted  ;D

and now its way past time I should be getting ready for work ...
I meant," said Ipslore bitterly, "what is there in this world that truly makes living worthwhile?" DEATH thought about it. "CATS," he said eventually, "CATS ARE NICE.

ldog

#147
Uhhhhhmmmm....ok, did not properly write what I figured but it still stands.
The formula I posted is meters per minute. Because our time unit is 1 minute.
So you would take the formula Tage posted and then multiply it by 60 to put the time into minutes, giving the 16.666666666666666666666666666667 I refered to.

Steve if you want to publicly challenge that figure, then like I said in my pm I think that is fine.
I am not going to get angry over this. I don't believe you are going to get angry. And as long as everyone else following along knows that no one is getting angry at each other then public argument (aka debate) is fine. As I said at the end of the last flamefest, as long as we all remain respectful of each other, then disagreement is no harm, no foul. We are here to have fun. No one needs to take any of this crap personal.

I am doing a lot of just plain braindumping (brainfarting ?) here and I hold to very little as fact. Even my observations so far are simply not enough to prove anything other than very simple things (the things done in very controlled experiments).

The speeds that they used, if we all agree that a tile is 16m and the time unit is 1 min and they are indeed the speed in Kph, then the speeds are give or take about 10/16 realistic reasonable speeds.
Even if I was wrong with my math, which I do not believe I was, then even still the above statement is true.

Prima guides rarely have correct technical information. They are meant to be playing guides not a technical reference to the game. I really do not put much stock into anything printed in them.
Since Prima has already gotten enough of my money over the years and I do not feel I have gotten my moneys worth, I am not about to send even another single dollar to them.

I stand behind what I said as a valid theory until someone mathematicly proves me wrong.

ldog

#148
Continuing our testing from where we left off last night;
We find our hero beset with traffic from all sides   ??? ...
I did 2 quick tests, about 5 years each (I saved each time though...so we have aged the city 10 years). For the first I set the trip starting cost for car to 2 (we started with .4) and for the second I went even higher to 3.
Remember this is not what I have been referring to as "Simulator L" this is the stock traffic controller stripped down even further at the moment. So we still have a max commute of 6.

What were the effects? Avg commute time rose slightly to about 3. I really wasn't expecting it to have much impact; I don't remember where but I do remember someone (I think it might have been Tarkus even, although it probably was Steve) saying that the trip penalty is subtracted from the max, not added to the actual trip. This would tend to confirm.

Also the pedestrian traffic rose quite a bit (we only have roads at the moment so there are only 2 choices, walk or drive). Mainly in the very congested areas where work was also near enough to walk to in 6 minutes :P or less (commercial and hospitals)

Next we'll do a bit more testing adding MT into the mix. As soon as I figure out which method(s) will give me the most bang for the buck (as far as ease of observation for comparison vs value of data gained) Stay tuned!

I decided to go with bus for the moment. I layed down a few central routes that covered about 1/3 of the city. I got some bus use. Pedestrians went up a bit as well of course. Nothing extreme. It did eliminate the flapping abandonment cycle I had on the far east side (the last row of residences would go to no job, abandon, refill constantly on their own). Effect on commute time was negligible. I let it run that way about 12-14 years. Then I covered the entire city with bus routes, and let it run the remainder of 20 years (so 6-8 years like this). Car shot down to less than half. Bus and peds were actually slightly above car. Congestion on all roadways went to green. Commute time dropped to 2.75

Heading back to the reader and dropping the starting trip cost down to 2.0 the city looked more normal. Car use rose back up, bus/ped dropped. Congestion did not return to pre-bus levels but there was a fair bit. This is with a TE cost of .274 (.96/3.5). Believe it or not I actually still found an instance of shortcutting across the bus station. Commute time actually dropped slightly back to 2.5 again. I let it run for 10 years. We are up to 80 years now.

Past my bedtime once again, but tomorrow I will try setting the starting trip cost back to its original value of .4 . I bet my bus use will go to hell. Anyone care to wager?

ldog

As expected going back to .4 trip starting cost by travel type for car made the bus use and walking in general plummet. There was still some bus use of course, and sims would take the bus quite far, they just were not going to do much walking.

For the next test I set everyone to MT preferred 100%. So by default now we have a trip starting cost for cars of 2.35 (.4 for being a car plus 1.95 because we prefer MT). As expected, traffic volumes and patterns went back to pretty much what they were before (with a starting trip cost of 2 for cars) Then I quickly tested again with the MT starting cost for car at 0.1 bus use plummeted, traffic went back to about what it was under fastest preferred.

For anyone who is wondering what the hell I am trying to do here; I am verifying that various parameters actually work as documented (or even work at all for some of them) and also that making changes to them produce the results one might reasonably expect them to

So in the current battery of tests to do still there is Max MT commute time and trip starting costs for car preferred (where we make walking even less desirable).
These are crude tests to determine working yes/no and make observations about the general effects. We will get to deciding how to make use of such parameters and refining the magnitude of changes and the cause/effect to them at a later date.

z

Quote from: ldog on November 19, 2009, 07:42:25 PM
For anyone who is wondering what the hell I am trying to do here; I am verifying that various parameters actually work as documented (or even work at all for some of them) and also that making changes to them produce the results one might reasonably expect them to

This is a very noble effort, and well worth doing.  However, I should simply mention my own experience here, which is that the same parameter can act in radically different ways depending on the settings of other parameters, and also which city is being run.  This led to some real surprises in some of the tests I've done.  Some parameters are obviously affected by this more than others; some are not affected at all.  Trip starting costs (both versions) were one of the worst offenders here, with their effects varying from huge to almost nonexistent, depending on the factors I mentioned.

Quote
So in the current battery of tests to do still there is Max MT commute time and trip starting costs for car preferred (where we make walking even less desirable).

I am very curious about Max MT commute time, as my limited tests were not able to give me a completely satisfactory answer here.  At this point, I believe it's the maximum length of any single MT segment, where "walking" counts as MT.  But I haven't had the chance to fully test this out, so I will be interested to see what you find.

b22rian

Quote from: ldog on November 15, 2009, 07:47:08 AM



Then take the fact that a bus holds how many people? 100? 150? 200? I really can't remember, I haven't rode a bus in many years. In NYC during peak times the buses get crammed as full as possible. When I lived in Seattle I also remember that most of the buses there were the accordian in the middle type...they were like double size buses. I've never seen one packed quite as full as a NYC bus but then if actual safety capacity was not routinely exceeded in NYC I would have to say it was a much higher capacity bus that Seattle had. So if 1 bus is like 100 cars, then how many buses do we expect on a given road during a commute trip in SC4? If we said like 20 buses, at 150 people per, that is 3000 people riding the bus through a given segment. While that would mean there are more bus passengers on the road than cars with low capacity settings, I don't think it is unrealistic, and I have yet to see my bus volume on a given segment go higher than car. Yeah, I said the R word again. Anyway, it is just something to think about, there is no correct or incorrect answer here. This I firmly believe is one of those things that comes down to personal preferences.



   Yup i agree with all this .. But i might point out as far as the game is concerned this is all based on perspective
and how one sees buses in the game.. For me when im doing my route queries , the bus number only indicates
to me how many sims are using the buses.. it has nothing to do at all with how many buses are on the road ways
Although if I had to guess i would assume most people playing seeing the bus numbers, think that means numbers
of buses.. which is fine.. this is all a game and we see thing in the perspective of how we want to see them really, and than play the game accordingly.. if you want to see that as meaning 1 bus carrying xxx amount of
sims , sure thats fine.. But that is not what it represents mathmatically, to the traffic sim..

Brian

ldog

Quote from: z on November 19, 2009, 11:50:33 PM
This is a very noble effort, and well worth doing.  However, I should simply mention my own experience here, which is that the same parameter can act in radically different ways depending on the settings of other parameters, and also which city is being run.  This led to some real surprises in some of the tests I've done.  Some parameters are obviously affected by this more than others; some are not affected at all.  Trip starting costs (both versions) were one of the worst offenders here, with their effects varying from huge to almost nonexistent, depending on the factors I mentioned.

This is very true. I am sure you would agree that the trip starting costs have a very strong relationship to the max commute time.

Quote from: b22rian on November 20, 2009, 04:22:27 AM
if you want to see that as meaning 1 bus carrying xxx amount of
sims , sure thats fine.. But that is not what it represents mathmatically, to the traffic sim..

LOL, no that is not at all what I meant though.
You are correct of course, the traffic sim simply sees the amount of sims using the bus, the same way it sees the amount of sims using the car.
Just like we don't have any carpooling mechanism we don't have a bus mechanism.
For instance it would be logical to assume that when we have multiple residents of the same building going to the same job that some of them would ride together.
There is just no way for the simulator to account for that, and once again if it was capable of it, it would really gain us nothing except added overhead.


z

Quote from: ldog on November 20, 2009, 09:44:19 AM
This is very true. I am sure you would agree that the trip starting costs have a very strong relationship to the max commute time.

Yes, I would agree, but most of my testing of these properties was done with the max commute time fixed at its current level in Simulator Z.  The effect of these properties still varied all over the place.

b22rian

Quote from: ldog on November 20, 2009, 09:44:19 AM

LOL, no that is not at all what I meant though.
You are correct of course, the traffic sim simply sees the amount of sims using the bus, the same way it sees the amount of sims using the car.
Just like we don't have any carpooling mechanism we don't have a bus mechanism.
For instance it would be logical to assume that when we have multiple residents of the same building going to the same job that some of them would ride together.
There is just no way for the simulator to account for that, and once again if it was capable of it, it would really gain us nothing except added overhead.



  yes, i 'm with you lenny..

Sorry, if I mis- represented your comments ..
I too wish that buses could be represented in the game in the way you are describing..
But it would require some major modifications in the transit modding of the game in order
to pull it off..

Brian

ldog

Quote from: z on November 20, 2009, 01:02:00 PM
Yes, I would agree, but most of my testing of these properties was done with the max commute time fixed at its current level in Simulator Z.  The effect of these properties still varied all over the place.

Yeah, I don't think if I ever figured out a formula for it that it would be anything simple. Probably something beyond my skills. We need Mott back  :'( .
Is there a mathematician in the house?
One thing I can say for sure, it is certainly not anything so simple as I originally thought, like if we double max commute then we double starting cost, etc.

Since I went out and got drunk last night I didn't do anymore experiments. Don't drink and mod! The sim you save might be your own.
I will pickup where I left off the other night (as soon as my head stops pounding)

Quote from: b22rian on November 20, 2009, 07:30:34 PM
  yes, i 'm with you lenny..

Sorry, if I mis- represented your comments ..
I too wish that buses could be represented in the game in the way you are describing..
But it would require some major modifications in the transit modding of the game in order
to pull it off..

Brian

No problem. Even though I got a laugh out of it, that is 1 more way someone could choose to look at it.
Like I said there are no right or wrong answers for this question. Even though our "grown-up" toys are much more sophisticated, we still need to make use of what made our simpler toys fun when we were children; imagination.
What I am aiming for here is some crude rule of thumb for myself as to how many bus sims vs how many car sims is "about right".
However you choose to see it, you have to answer that question for yourself before you can determine if your simulator is functioning properly or not (at least as far as bus usage).

I think it would be an extreme strain on the traffic simulator if it had to actually account for vehicles and fill them with passengers.
It would likely be something much more complicated than just slipping in a flag that says "X passengers = Y bus"
The game probably would have shipped with a PH of 9.0 instead of 0.9  ;)
I'm sure Steve could explain it a lot better than I could.

ldog

#156
Picking up from where we left off on Thursday night, just thought of something. I didn't verify if the plain old trip starting cost affects MT preferred or if it is only for fastest method preferred.
So I'm going to go ahead and crank it up to 4.4
I ran the sim for several years. No change whatsoever. Opened up the reader just to be sure I had saved the change (I had).
So it would appear starting trip cost by travel type applies only to fastest method preferred travel type. So we can be completely sure later I will run the same tests again with 100% car preferred.

Now I have gone ahead and changed everyone to 100% car preferred.
For the first test, we leave the other parameters alone. So the starting trip cost for walking under car preferred is 0.1
As expected car use shot through the roof. Pedestrian use exactly matched bus use (I had to uncheck bus volume to see ped volume). So they walked only to from bus stops and minimally at that.
Commute times went up, since we made the traffic worse again.
I let the city run a good 12 years (bringing us up to the year 100) and saved it. I wanted to get it good and stable. Interestingly enough I got the rolling abandonment (or flapping as I refer to it) on the East side again. No job zot. Traffic sim run. Abandon. Traffic sim run. Rebuild. Traffic sim run. Find a job. Traffic sim run. No job zot. Rinse and repeat. So commute time, volume, population all jittered slightly through the whole period but it is otherwise as stable as it can be.

catty

#157
Quote from: z on November 20, 2009, 01:02:00 PM
Yes, I would agree, but most of my testing of these properties was done with the max commute time fixed at its current level in Simulator Z.  The effect of these properties still varied all over the place.

Don't know if this is relevant or not, but in this link

http://www.gamefaqs.com/computer/doswin/file/561176/23133

which is on "Transportation in SimCity 4" he says in it

QuoteThe commute works in adding up the total travel times of all the tiles used for a trip, taking into account multipliers for traffic intensity and busy intersections. The total is then rounded up to the nearest integer in minutes, and put in average with the rest of the totals that were drawn in that cycle.

I've read something similar in another topic, but can't now find the link to it  :-[
I meant," said Ipslore bitterly, "what is there in this world that truly makes living worthwhile?" DEATH thought about it. "CATS," he said eventually, "CATS ARE NICE.

z

I don't know about that particular quote, but I took a quick look through the rest of the document and found a number errors in it (some major).  So I wouldn't rely on what's said there without confirmation.

ldog

Ditto what Z said. It was pre-rh, I think it was also pre anyone actually hexing the traffic controller.

Quote: As of Patch 2, there has not been any real algorithm to make your sim-motorists
understand which route is the "quickest."

Completely incorrect. As we all now know (due to many other peoples efforts) that it just worked so poorly with the .9 ph that it sure seemed to be true.
A lot of misinformation through the whole thing. It does show how far we've come.

I am not sure if he is speaking of the actual commute time or the display graph either. I am also not going to try to evaluate the correctness of either right now.

Does bring up a relevant point though.
In the last several posts where I am speaking about several different vars that have the words commute and time in them and I hope I have been careful enough in differentiating (my $5 word for the day) which one I am talking about when. I will go through and edit as needed later.

More test schtuff

I went ahead and cranked the old "trip starting cost by travel type" to 4.4 and running for a few years, I observed no difference whatsoever.
So I think we can conclude that it really should read "trip starting cost by travel type for fastest method preferred"

Moving along I put that back the way it was and I set the pedestrians starting cost under car preferred up to 4.1 and go run it a few years, which as expected all but terminated the bus.

So I have verified that all 3 are working, and by inference also that our travel type preferences all work.
Steve, someone, anyone...please let me know if y'all think I missed something.

Next for the fun stuff. WTH "Max mass transit strategy trip length" really means.