Transport Tycoon Forums

First, just so there's no confusion:

I know this has been suggested before. I know it is very difficult, and on the borderline-impossible list. I realize that it's unlikely Patchman would do it anytime soon, if ever. This is not intended as a "nag" post. So, in that sense, it's not a serious suggestion. What I'm intending is something more like a thought exercise as to how such a feature would work. Maybe it would come to be in OTTD, or in some other project, or maybe just help direct development of other things.

What I mean by "passenger destinations" is this. When a passenger appears, generated by a city, industry, or whatever, the passenger doesn't just magically show up at the nearest station and get on whatever vehicle happens by. Instead, the individual passenger has some degree of intelligence, and most specifically, has a designated goal in mind. That goal may be a specific city; or may be some sort of feature (i.e. "tourist attraction", "bank", in which any building meeting that goal will do); or may be some other station in that same city (local traffic).

The passenger, on arriving at a station, would peruse the schedule - that is, look at the vehicles that stop at that station and see where they go. If a vehicle provides direct service to their destination, they would wait until it came by; otherwise, they would consider connecting services. I would expect that there would be a limit on these, so that if a passenger could not figure out a way to get to their destination within, say, 4 transfers, they would give up and go home. That might be displayed in the station window as a failed destination, so the player could see that there was unmet demand, and maybe start a new service to meet it.

So, on arriving at a station, the vehicle wouldn't completely unload, as now. Only the passengers wanting to get to that destination, or needing to change there to another vehicle, would get off, and the rest would stay on.

This would allow a more realistic passenger network, involving local trains as well as express trains, and connecting services.

What would be the consequences of this change? What elements would need to be considered? What would be involved as far as displays, GUIs, etc.? How profoundly would the gameplay be affected? For the better, or for the worse?

Thoughts?

I voted for the fourth option, that it's an OK idea, but that I'd rather see the time spent on something else. Specifically, CPU time. It seems to me that all of the hundreds of passengers that can be generated each month later in some games all trying to figure out and remember a destination would drag down the CPU and memory.

Good idea ... and as I mentioned in OTTD.

Having to provide a service for passengers, that want to go to a destination of their choice, would be a great addition to the game.

Have a look at Traffic Giant. BTW, this game has nothing to do with 'Transport Giant', although it comes from the same company.

In Traffic Giant, passengers choose their destination. They will only board a bus, tram or lightrail if it gets them to their destination. You do not need to provide a direct service, passengers will transfer from one service to another as long as it gets them to their destination. If you do not provide an adequate service, they will take their own motor car and clogg the streets.

That is just what is missing in TTDPatch.

Flamelord wrote:... CPU time and memory ...

May be a group of passengers think the same; i.e a percentage wants to commute to a town nearby, another group wants to travel to the other side of the map. That would limit the thinking of each individual.

I would figure that there would be some sort of percentages.

25% would want to go to some other station in the same town (in-town traffic). Maybe not until the town reaches a certain size.

25% would want to go to the nearest 4 adjacent towns (local traffic).

25% would want to go to the largest city on the board (or maybe the largest city within X squares, or maybe split between two or three largest cities if there are a couple of really big ones.) This represents the excess traffic to really large cities and capitals just because they are big, like New York, LA, London, Tokyo, etc.

25% would be random traffic to Everywhere Else. This encourages the network effects.

It would be good if passengers had the ability to transfer between companies, and to do an in-town self-transfer. I.e. they can arrive at a station belonging to one company, and transfer themselves to a station belonging to another company to continue their journey. Just like it used to be with changing airlines, or changing rail companies in London via taxi or tube between stations. This would encourage you to build to a city mostly served by a competitor, because you could collect onward traffic that wants to go to a city you serve, but your competitor doesn't. And vice versa of course.

Ah. Well, if they traveled as groups, I imagine it might add some interest without excessively taxing your CPU and/or memory.

One problem would be having feedback from passengers indicating where they wanted to go.

At the moment, passengers appear at a station and want to travel somewhere but it doesn't matter if they are moved to the next bus stop or to the other side of the map. Basically, the further we carry them, the more profitable it is for us.

Now, for the 'destination' scheme, if they're stacking up at a station we need to know where they want to go so we can consider whether to make a new connection. To a great extent subsidies fulfill this demand led service. If we're so inclined to create this service then fine.

Another way to engineer this is to split the map into 4 sectors, NW,NE,SE,SW and passengers randomly choose a destination sector. They arrive when they debark at a station in the correct sector. Problems arise there though, if you have a train carriage with 40 passengers, do all 40 want to go to the same destination sector? How can we tell??

This would be a great feature, and would create finally a true reason why we should build both express and local traffic lines.

One way to solve the problem 3iff pointed out, would be to make the passengers' destination appear on the station box and on the vehicle cargo description; passengers could be clustered by destination (e.g. 30 passengers to town X, 10 passengers to factory Y) so that we would know where they would get off. And if a vehicle already carrying passengers to a town stopped at a station where other passengers going to the same town were waiting, (for instance on a local railway line heading to a major city) these waiting passengers would just board it, as long as there were still room available, and be added to the cluster corresponding to their destination. Such a scheme, for instance, could model the commuting lines linking the villages at the outskirts of the major cities to the center of the city itself.

Connecting passengers would drop off at a station and be added to the cluster going to the same destination in the station's waiting list.

It might be very hard to do, but I'm sure this would be one of the features which, when done, would be a major improvement in the game, giving it a totally new flair and forcing us to study and optimize our networks even better. It would also add a lot of realism. So I guess that, even if it might be a hell of a work, it would be time well spent and definitely worth the effort.

O/c this could apply to mail too. Mail doesn't just get delivered anywhere, it always has a destination

Snail wrote:This would be a great feature, and would create finally a true reason why we should build both express and local traffic lines.

Haven't got anything as such to add to the idea, but just wanted to comment that you can still build reaslistic passanger networks in TTD the way it is right now... The passangers just get off on the next station

The best thing about this would be that routes with lots of small stops a short distance from eachother would become very profitable! Currently, each passanger just hops to the next one, and you get very little money. With passanger destinations, a route with lots of stops a short distance from eachother might end up more profitable than a long distance ICE train across the map

Also, think how big the trains would have to be! If the passangers didn't get off on the first station, passanger trains would have to be a lot longer... like they are in real life.

It's a dammed pity that coding this into TTD Patch is a near impossiblity

The first thing needed before writing a single line of code would be a detailed requirement specification because "adding this feature" would be a really challenging and complicated task - possibly resulting into a totally different game.

regards
Michael

I 100% agree with you, michael, and that's really the reason for this thread! We can bat things back and forth and come up with the spec, and then see where (if anywhere) it goes.

I agree it would be almost a different game. But it wouldn't be THAT different, because it would only affect passenger services. I view it as another evolution in complexity, kind of like PBS signals.

Concerning profitability, I think overall passenger services would make more money, because the average distance travelled by each passenger would be longer. I don't think that short services would be more profitable than long though, because the payment algorithm would be the same - distance traveled vs. time taken to travel it. Of course, the passenger would pay separately for each leg, I expect.

Now, you could maybe game the system by intentionally making passengers travel long distances out of their way, but maybe there's a way to discourage that. I suppose passengers, when considering which route to take, could try to count up the number of squares they would be travelling, divided by the average speed of the vehicles, to figure out the quickest way. No way to figure out wait times, though, that I can see.

Doesn't Simutrans do that? There might be some kind of specification there...

In addition, it has been discussed for OTTD <cough>, e.g.:

http://www.tt-forums.net/viewtopic.php?t=9201
http://www.tt-forums.net/viewtopic.php?t=2193838

regards
Michael

3iff wrote:One problem would be having feedback from passengers indicating where they wanted to go.

You would start with the town/city (info window); where the player can find out what the inhabitants intentions are. e.g. there it may say 'Commuters to town A (15 % or 75 per month)'; 'Long Dist Pax to City B (35 % or 175 per month) etc

After establishing a service to town A, 15 % of passengers created would show up at that station. The station info would then show 'n passengers waiting for town A'. If the service provided is inadequate or ceases, waiting passengers would disappear slowly.

I suppose this ought to be linked to the subidies if it ever happens. I'm in a hurry, so I'll have to explain later.

OzTransLtd wrote: You would start with the town/city (info window); where the player can find out what the inhabitants intentions are. e.g. there it may say 'Commuters to town A (15 % or 75 per month)'; 'Long Dist Pax to City B (35 % or 175 per month) etc

That's a good way of doing it. One slight problem would be space to include this in an efficient way. I would assume that passengers would want to travel from a city to any other city on the map. With the minimum of 12 cities, that would be 11 other lines to list. With lots of cities, that would be 25-30 or more (not sure of the max cities permitted but did I see 70 mentioned?).

I suppose a city info window could list the top 2-3 destinations to make info meaningful...after all, who wants to know that 3 passengers/month want to travel to Hafingford when 750/month want to go to Rarington...

I'm with Flamelord though when I keep coming back to associating this idea with subsidies...I just can't formulate how I would do it.

I would say, that you could have some sort of a trigger. When a particular city has more than X passengers that want to go to another city, and there is no service to do it, then a subsidy will be offered. That seems realistic to me.

3iff wrote:That's a good way of doing it. One slight problem would be space to include this in an efficient way. I would assume that passengers would want to travel from a city to any other city on the map. With the minimum of 12 cities, that would be 11 other lines to list. With lots of cities, that would be 25-30 or more (not sure of the max cities permitted but did I see 70 mentioned?)...

I would say the maximum number of options for each town/city should be limited, say 16 (a round number); i.e. all passengers created each month are split into a number of groups. In the early years you may only have 4 or 5 of them. That arrangement is periodically reviewed as the town growths. Now each group intends to travel somewhere, say commuting to nearby towns, or travelling long distance to other small towns or cities. You may also have to 'collect' some passengers, using local bus services within a town to get them to your station, located at the edge of the town.

You would not want to have a lot of groups containing a handful of passengers wanting to go everywhere. You get the spread to everywhere if you get those passengers (groups) from, say 4 or 5 small towns to a nearby transport hub.

Subsidies could of course play a vital part; especially if a large group of passengers gets neglected. You'll get your nose pointed where services are required.

Personally, I don't think this is a great idea. If it were implemented poorly, it would be irritating and no-one would use it. If it were implemented well, it would be so transparent that it wouldn't really make a lot of difference to the gameplay.

I remember proposing this idea almost ten years ago through Usenet or another Tycoon web page.

I would think many people would not appreciate this idea because most people prefer creating large trains to haul goods and raw materials while passenger services are secondary. Also, because the total number of passengers are broken into percentages, having one train taking all people away in one swoop is easier than perhaps using a lesser service type like buses spreading the load. Anyway, let's discuss the potential for this idea.

I am not a math wizard so the formulas and number crunching shall be left to another. Town A has a population of 300 people with 50 people desiring to travel. Town B has a population of 500 people and is 10 squares away. Town C has a population of 150 people and is 15 squares away. Town D has a population of 350 people and is 10 squares away. Thus among the three potential towns to travel, there are 1000 people in population. Therefore, using town population / total population, B = 50%, C = 15%, and D = 35%; and these percentages multiplied for the original 50 travelers means approximately 25 would go to Town B, 7.5 would go to Town C, and 17.5 would go to Town D.

But now the distance should be a factor. Both Towns B and D are 10 squares away while Town C is 15 squares away. Since the two towns with 10 squares are the shortest distance, all 100% of the potential travelers would head to those towns. Yet with Town C, (15 squares - 10 squares) / 10 squares * 100% = 50% of its potential travelers not making the trip. Thus 50% of the rounded 7.5 travelers would be 4 travelers going elsewhere and with the two other towns being equal distance would have these other travelers divided evenly. Thus Town B would have (25 + 2), Town C would have around 4, and Town D would have (18 + 2) of those original travelers from Town A.

Now you just need to have a route set-up for them to get there.

A general idea as a start for further discussion and possibly for another to work out the greater details with number crunching and perhaps a random factor added within the formulas too.
>>>>>Axlrose - ...<<<<<

what is the differince between Simutrans and openttd that they can't make a system and in openttd it stay a dream forever.

It is really a shame for openttd.