I'm Saddened... (Board Games)


I know this should probably go in the specific Magic thread, but considering it’s more of a look at Magic’s history, and given Magic’s influence on board/card gaming in general, I thought that posting here would be appropriate:


I saw that on Reddit yesterday. I’m not sure how it is for an outsider but as a decade long entrenched player it was a fantastic read.


As someone who never got into Magic and has only a cursory knowledge of it, I thought it was also fantastic.

What I found really interesting, and I’m not sure if this is just the New Yorker being generous to WotC/Magic or not, but how Magic was/is still trying to grapple with representation and inclusivity. As primarily a board game player, I’ve been peripherally around Magic tournaments at gaming stores, and the majority of the players seem to be white straight men, but maybe that was just the stores I played in. I like that at the higher tournament levels, Magic is really trying to focus on welcoming all players, regardless of race, gender, or sexual orientation.


The tournament scene isn’t as diverse as the kitchen table casual scene but the important part is that they strive to make people feel included. They don’t control who the player base is but they try to make the art and story of the game as inclusive as possible and as such though the game is dominate by white men even the tournament scene is remarkably inclusive of people regardless of who they are.


I’ve read a lot of board game strategy articles, and don’t mind that they’re a shortcut to understanding.

I’ve read a lot of 18xx articles and the game still hovers juuuust out of reach. I don’t really “know” 1846 after 20 games, and now I’m playing 1889 and 1860 and I’ve pretty much waded into the ocean.

I was able to memorize Terra Mystica openings pretty quickly. With 18xx the answer to so many things is “it depends”

Some of my favorite games work the same way… Race for the Galaxy, Tigris & Euphrates, but in those games you’re contending with probability and your opponents in equal measure. 18xx is, to paraphrase an old Geeknights, brain bashing against brain. Sometimes you try something weird and it works because no one knows what to do and somehow you come out on top.

It really is a remarkable system and the only downside is it takes 3 hours and you want a calculator or the ability to do arithmetic quickly.


Watching this AMA


I treat 18xx the way I would a wargame. I’m in it for the experience. I’m trying to win. But I’m approaching it the way I approach real-world problems with real-world heuristics.

Normal tabletop, I approach with finely focused heuristics that make a lot of assumptions about rules of engagement, parameters of interaction, time/action economy, etc… I make choices based on specific calculations, exact odds , etc…

e.g., I consider a turn successful in Power Grid if I spend exactly all of my dollars. I consider it a failure if I’m off by a single dollar in either direction. I take risks based on certainty of outcome.

18xx/wargames, I use looser heuristics based on plausibility of outcomes.


The wargame comparison half-works for me. 18xx games aren’t necessarily a simulation… a system exists and you have to extract money from it, but unlike most wargames or John Company (which is also a system you need to extract money from, sort of) the system is deterministic.

Where it gets really fascinating is the differences between games. In 1846, capital is constantly pouring in, in greater and greater amounts, and you’re trying to ride that wave and push the other players out of the way. In 1889, share value and even capital can be destroyed at a whim during a stock round. The trick is dunking your opponents instead of yourself.

At any given moment, you can look ahead and say “if no one provided any further inputs to this game, player X would win”. The trick is in finding the right buttons to push that will not only make you that player, but make it difficult for your opponents to become that player.

New players often miss some of the available inputs. I sure as hell do, 2 years later. Rusting trains is big, but only if you’re losing. In that 1889 we played, I rusted a bunch of 4s but the steps I needed to take to do so probably hurt me even more.

I was also making so much money from privates early that I could have happily sat by for one more set and let you operate share companies while I pulled even further ahead. Scott buying up to two shares in pink gave me control over $700 to buy trains with no obligation attached. I robbed the treasury to buy my privates and immediately sold those shares to create another $900 in a fresh company, which rusted those trains that now belonged to Scott. That should have been the killing blow but I fucked up by withholding twice to buy a Diesel, losing a bunch of share value and not earning very much in dividends.

Edit: Even if a company is going to “lose” a train it’s sometimes more valuable to continue to pay out dividends and then fund the train they need out of your pocket, due to the stock value increases you get from paying out.


Yes. But in a way that I am unwilling to fully engage with and calculate. So I treat it like a real world situation. I focus on a long term goal, a short term goal that feeds into that long term goal, and a set of actions that could plausibly effect my short term goal.

What I don’t do (though I could in theory) is look at specific prices in detail to determine exactly what to do. I never bother with the Power Grid or Advance Wars level of “if I need to spend $x for y in two turns, I can spend $40 on a token now with no consequences due to z happening next turn.”


That’s fair. I also don’t worry myself with that detailed degree of tactical maneuvering or tile laying, outside of maybe one turn of lookahead (and even then, it’s mostly how I’m going to get trains into companies). I have an idea of what I’m trying to do strategically, and a small but growing reservoir of experience to inform the “how”.


What fascinates me is how interesting all of this ends up being in practice.


Because it is deterministic I think we could actually make a AI that plays some 18XX games nearly, (or absolutely?) optimally. The problem is, as Rym said, there is no way a human can make all those calculations at the table if they are pressed for time and don’t have a calculator. Maybe a genius player could do it?

I think a major problem I have with all the 18XX games is that I never remember all the possible things I can do. In the one we played this weekend I didn’t remember that upgrading tracks was an option until I saw somebody else do it. I did manage to finally remember that buying a new company was an option, though! I just still have to learn how to leverage that into success.

In multiple games now I have had failures due to trains being deprecated, but not having enough money in the train company to buy new trains. I recognized this and tried really hard to not have that problem. I managed to buy new trains right up until that final rusting in our latest game. One more withholding without rusting and I would have bought the diesel. I feel if I can figure this out, I can be successful, even if I don’t win.


This is probably correct in 2p, similar to how heads-up texas hold’em is effectively solved. Train an AI on a specific set of rules and watch them go. As the player count increases, I expect the chaos introduced by 2-5 other players colluding with one another to bend the game in different directions makes multiplayer significantly more difficult to solve than, say, Race for the Galaxy.

(I bring up collusion to cover games like 6p 1830, where no single player has enough capital to float a company in SR1. You would have to negotiate to survive in those games, and it can be true to a lesser degree in smaller games)


I’m still waiting for someone to create a truly modern, clean implementation of 18xx that can be played aync in a browser. Existing solutions are relatively archaic, similar to PBEM Civ without Giant Multiplayer Robot.


I bet I a one-page heuristic expert system could get you most of the way there. If this, do this.


How about you make one. Pick a year that hasn’t been used yet, and go for it.

  1. Amtrak shit.


I can’t find it now but I recall a thread discussing how there are more than 100 18xx games in various states of publication, which means at least some of them overlap or break the naming convention. Thematic names (Poseidon), different centuries (2038: Tycoons of the Asteroid Belt) letters (18AL) etc.

Making a new 18xx game worth playing is really hard. Making a boring set of rules is very easy. And your game will only ever get played by people who deeply care if it’s interesting or not - there are so many variables beyond the map. Full or partial capitalization? How much starting money? What are the trains like? What is the tile mix (all the yellow cities in 1830 are straight track)? How large is the bank? All the way to sacred cows like how 1860 lets you sell a public company into receivership.


Steal from the (forgotten) masters.


Maybe it’s hexadecimal and they forgot to tell anyone. So you can go up to 18FF before running out.