Home

Rails Across America

Sat, 10/18/2008 - 00:41 — Rick-Holzgrafe
Rails Across America Board

In Rails Across America, players compete to build the most valuable railroad network. A network's value is comprised of the number of towns and cities it connects, the number of railroad stations built in cities, and the amount of money the player has earned.

The board shows a diagrammatic map of America: a network of lines ("links") meeting at junctions. (See the attached image.) Six of the junctions are major cities, the rest are towns. There are four grades of links, varying from short/easy to long/difficult. The more difficult a link is, the more it costs to build, but the more money a player makes when delivering goods over that link.

There are four kinds of goods, represented by cubes in four colors. During setup, one random cube is placed on each junction. Each junction is also marked with one of the four goods colors, which indicates that that city or town is a potential destination for that kind of good.

Players build links by paying money, according to the link's grade. A built link is marked by placing a token on the link in the player's color (like a Catan road, or a Railroad Tycoon choo-choo). The first player to build a given link pays the bank. A second player may later build the same link by paying the cost to the first player instead of the bank, and then placing his link marker alongside the first player's. No more than two players may build any given link. All of a player's links must be connected; discontiguous building is not allowed.

Deliveries are made by moving a goods cube from its starting junction to a destination junction whose color matches the cube. The active player may choose any route, with these restrictions: All links must belong to the active player, the route may not loop or return upon itself, and the cube may travel over at most seven links. You may move a cube past a junction of matching color in order to reach a further destination. The player receives money from the bank for each link in the route, according to the link's grade.

When a cube is delivered, it is removed from play and its starting junction is left empty. A player can "replenish" (place one new random cube on each empty junction in his own network) by paying a cost to the bank.

Each of the six major cities has a stack of Delivery tokens, marked with a monetary value and with the city's name. When a player makes a delivery to a city, he may take the most valuable remaining Delivery token for that city and keep it.

Each of the six cities has three Railroad Station tokens, each worth a different amount of VPs. When a player builds a Railroad Station at a particular city, he takes the highest-value remaining Railroad Station token for that city. Building a Railroad Station costs a lot of money, but Delivery tokens for that City may be turned in as part of the payment. (A Delivery token is essentially a discount on building a Railroad Station at that City.) Spent Delivery tokens are removed from the game.

At the end of the game, VPs are awarded for the number of cities and towns in each player's network, and for the amount of money the player has left. For cities, the VP rewards are triangular: 1, 3, 6, 10, 15, 21. This is to encourage vigorous competition to build a wide and comprehensive network.

Turn order goes clockwise. On each turn, the active player takes two actions. He may take the same action twice, if desired. The actions are:

* Build one link
* Build one Railroad Station
* Deliver one Goods Cube
* Replenish
* Take a Bank Loan (get money, but lose VPs at end of game)

End of game is triggered by running out of cubes during a Replenish action. Play continues to the end of the round, so that all players get the same number of turns. Final VPs are calculated, and most VPs wins.

For a discussion of why I'm designing this game, see my blog.

Comments

Wed, 11/26/2008 - 03:58 — Rick-Holzgrafe

Balance

Nando recently asked me for an update, so here it is. The simulator has been very useful, and very illuminating. It's helped me to settle on a number of different tunable parameters, and helped me adjust the map so that it provides a reasonably balanced four-player game.

The bad news is that it's also shown that the game is seriously unbalanced for three and five players, and the only way to fix it is by redrawing the map. I fear that the game will need a different map for each number of players I want to support. That kinda sucks.

If I'm lucky, I may get away with a two-sided board: four-player on one side, five-player on the other, and a designated subset of the five-player board for three players.

Those interested can read a bit more detail in my blog.

My other concern is that the game is simply too unoriginal. It's not exactly like any other game, but it's still a fairly generic build-network-make-deliveries train game. I hope to hold some real-live-people playtests soon, and I can ask the players what they think. If they're unimpressed, I may drop the project.

Mon, 11/17/2008 - 20:33 — Rick-Holzgrafe

Simulation

(Edit: I've just discovered that making a tiny edit to an older post is treated the same as a new post. I wish it wouldn't do that; I just wanted to make a little correction to my grammar, not make everybody look again. This post about "Simulation" first appeared on Friday Nov 14 and has not changed substantially.)

Daunted by the task of balancing and tuning this design, I've done something radical: I built a software simulator for Rails Across America. This gadget is written in Java, and its job is to let me twiddle the game's tuning knobs (costs, payouts, number of players and cubes and so on), and then play a few thousand simulated games and show me the results.

When I do solo playtests, I pretend to be three or four or five different players, and play out an entire game to see how it goes. That's what the simulator does, using an AI module to make the decisions.

In this case, "AI" means "Artificial Idiot". It's not as bright as an experienced human player would be. But I've tuned it to where it plays a credible beginner's game. And it can play 10,000 games in under five minutes, so I can try a lot of different variations in a single evening.

Here's some sample output:

Stats: 1000 games, 34 ties.
Most rounds: 27, Average: 22.667
Lowest cash: -387, Average: -116.903
Wins by order: 264, 270, 254, 212
Average scores by order: 24.584, 24.964, 24.332, 23.488
Average major cities by order: 5.031, 5.008, 4.99, 5.001
Average stations by order: 3.887, 3.97, 4.023, 3.906
Average switchyards by order: 0.781, 0.78, 0.754, 0.702
Average loans by order: 0.39, 0.407, 0.408, 0.423
Wins by start city: Denver:165 Chicago:25 Los Angeles:277 Miami:167 New York:103 San Francisco:263
Elapsed time: 23.873 secs

In this listing, "by order" means that the numbers are listed in order for the 1st, 2nd, 3rd, and 4th player. (You can see that there's a small advantage to going early in the turn order.)

"Lowest Cash" helps me track the flow of money in the game. I use a little programming trick: I start the bank at zero dollars, and let it go negative as money is given to the players. In the listing, "Lowest cash: -387" means that in all 1000 games, $387 was the most money that was ever in circulation. The average was about $117. That means that in a real game, the bank should start with $300 or $400 dollars in it, so the players won't run out of currency. (So how exceptional is that $387? I don't know yet. I need to track some of these values in more detail, which is a feature I haven't gotten to yet.)

I've also posted about this at my blog.

Wed, 10/22/2008 - 21:22 — Rick-Holzgrafe

New Scoring System

I've played several simulated multi-player games now, each time with a different scoring system, to see how they play out. The latest was the best yet, and seemed to produce an entertaining and competitive game.

The major changes were:

* Removed the scoring for connecting major cities.
* Changed the way in which Railroad Stations are scored.
* Added "Goal Cards".

Railroad Stations are now scored similarly to the way that connections to major cities were: the more you have, the more each new one is worth. There is a limit of two stations per major city, so there is still tension in deciding whether to build now, or wait until it might be cheaper.

Goal Cards are similar to some of the Operations cards in Railroad Tycoon: each specifies a goal to be achieved, and a VP reward for achieving it. But they work more like the Public Objectives in Twilight Imperium 3: any player can score any goal at any time, no matter who else has already scored that goal. Only some of the Goal Cards are used in any game, so each game is different.

Although there are no longer built-in rewards simply for connecting major cities, most of the goal cards specify connections, some specifically (e.g. "New York to Los Angeles") and some more generally ("any four green towns"). Some of these do allow scoring for connecting certain major cities.

I suspect that there is an advantage for the players who start in the west. The connections there are more expensive, but pay off better: and you pay for them only once, but profit from deliveries many times. That may require tuning.

If I can convince myself that this isn't too much of a Railroad Tycoon clone, there might be a game here.

Sun, 10/19/2008 - 04:47 — sedjtroll

Nice!

I like the sound of this, Rick. It does sound fairly similar to other rail games, but I guess to an extent all rail games share several similar aspects. It's clearly inspired by Railroad Tycoon I think, and for some reason it reminds me of Steel Driver (which I haven't actually seen in action yet). But it sounds decent, and it sounds like the game play could be original enough.

I like that 2 people can build any given route, and that the 2nd person to do so pays the first - so getting a route first means you get it for free basically, unless no one else gets it (how likely is that to happen?)

The Replenish action reminds me of Hansa, and I wonder if it might be a good idea to make it even more like Hansa's replenish - replenishing the whole board and not just your routes. Though I'll note that when you replenish your own routes, another player could still deliver those cubes (since multiple routes go into towns and cities). In fact, I could see a situation in which you're damned if you replenish - all the cubes could be delivered before you go again, and damned if you don't - you have no deliveries to make. This could potentially pose a problem. I'll also note that in Hansa, replenishing isn't a turn action, it's something you can do at the beginning of your turn for a cost. That might also be something to consider (if the stalemate thing occurs).

Good work, keep us posted!

- Seth

Sun, 10/19/2008 - 15:41 — Rick-Holzgrafe

sedjtroll wrote:I like the

sedjtroll wrote:
I like the sound of this, Rick. It does sound fairly similar to other rail games, but I guess to an extent all rail games share several similar aspects.

True, and yet I'd like to inject a little more originality into it. Not sure how yet. I might just set this aside for a while (I need to work on Heir and Regent anyway) and see what bubbles up after a couple of months of ignoring it.

Quote:
It's clearly inspired by Railroad Tycoon I think, and for some reason it reminds me of Steel Driver (which I haven't actually seen in action yet). But it sounds decent, and it sounds like the game play could be original enough.

Definitely Railroad Tycoon -- see my blog entry. The system of colored goods cubes to match colored cities is RRT/AoS. I haven't played Steel Driver yet either, although I've read the rules and it looks cool. My map resembles Steel Driver's in that the potential links and their building costs are drawn on the board. (This is also true of Ticket to Ride and Power Grid, no doubt among others.) The payoff for deliveries is more like the crayon rail games: you make a delivery, you get immediate cash. The payoff-per-link is not exactly like any game I can call to mind, but perhaps most similar to RRT.

Quote:
I like that 2 people can build any given route, and that the 2nd person to do so pays the first - so getting a route first means you get it for free basically, unless no one else gets it (how likely is that to happen?)

So far it seems that most links aren't doubled, but some strategic ones are. Connections to the major cities are crucial, so the early builders of those links can usually count on some "leased rail" income later in the game. There is usually also some other "leased rail" on routes between the major cities, but as there are several ways to skin that cat, you can't count on that income—your opponents might go a different way.

On the whole, not too much money changes hands using this mechanism. But a major part of the gameplay is alternating deliveries with builds, making money so that you can spend it. Getting a little unexpected (or expected) income between turns can speed up your development. Also money is worth VPs at the end of the game, and the average link cost works out to about one VP. So another way of looking at it is that you're giving the other guy a VP when you double his rail. It gives the players more to think about.

Quote:
The Replenish action reminds me of Hansa, and I wonder if it might be a good idea to make it even more like Hansa's replenish - replenishing the whole board and not just your routes. Though I'll note that when you replenish your own routes, another player could still deliver those cubes (since multiple routes go into towns and cities). In fact, I could see a situation in which you're damned if you replenish - all the cubes could be delivered before you go again, and damned if you don't - you have no deliveries to make. This could potentially pose a problem. I'll also note that in Hansa, replenishing isn't a turn action, it's something you can do at the beginning of your turn for a cost. That might also be something to consider (if the stalemate thing occurs).

The Replenish action is lifted directly from Hansa. (You see what I mean about there not being much originality in this design!) Your analysis is essentially correct. Overlapping networks does mean that the cubes you replenish can be stolen by other players. For this reason it's best to replenish as the first of your two actions per turn; that way you can be sure at least to deliver the juiciest cube before anyone else can. OTOH, it's rare for all of your junctions to be shared, and rare for all other players (in a 4p game, anyway) to have a shot at your cubes. You can expect to lose some cubes, but it's very unlikely that all of your cubes can be stolen before your next turn.

Also, by the time the players' networks are that deeply intertwined, deliveries are long and lucrative. Getting even one decent delivery, even minus the cost of replenishment, should pay you enough to make a couple more builds, or perhaps purchase a station at a major city. So while there is definitely some worry about benefitting your opponents when you replenish, there's still plenty of reason to do it when you need to.

I think it's best that Replenish is a distinct action. It increases the brinkmanship: "if I can hold out just one more turn, I think Green will have to replenish, and then maybe I won't have to spend the money and the action myself." I like games in which actions are a scarce resource.

I don't think the whole board should be replenished. That makes the replenishment schedule too chunky, if you know what I mean. One replenishment of the whole board would be roughly equivalent to maybe three or four replenishments of individual networks. I'd like the action to be more frequent and in smaller chunks than that. Remember, in Hansa the board is not very large, and everybody visits every city frequently: not true in a rail game where you have to build connections before you can go places.

Thanks for the input, Seth! If I get motivated, I might post the actual rules soon. Although I really need to spend some time on Heir and Regent!

Syndicate content