Arabella Initial Thoughts

Played Arabella tonight. My initial thought was “This was like asking Chat GPT to make an 18xx roll and write, but not the most recent ones, like 3.5 or 4.0, but ChatGPT 1 that was prone to hallucination.”

There are certainly aspects that sound like 18xx. You build track, buy shares, buy trains (and carriages) run trains, pay dividends, trains “rust” (but not really). BUT even ignoring that this is a roll and write (for a system that mostly appeals because of its minimal luck) there are issues:

Arabella doesn’t capture the cadence of 18xx — Early track builds are free, and then get expensive. So the first few turns are build build build! You can buy stock only when allowed by dice. Once you get a good run, you can use any pair (or three of a kind) to run for good value. So my game was something like build build build, take cash a few times, buy a few shares, buy a train, run, buy a train, run, extend the track and now I had a nice income on my run, so running ~10 times would end the game and simultaneously get me lots of objective cards (which are the primary source of VP) so run whenever a pair showed up.

Trains never rusted — (See this thread for others complaining as well).

Weird stock things — Each company only has four shares … I guess the idea is that the president owns 60% and the other 40% are owned by others, but you can never lose “your company” and can only buy shares in the other players. Except … the president can buy shares back from other players into the company, denying VPs. Um, ok?

There is some snowballing, like in 18xx, but this is a dice game — So when I rolled an early five of a kind (allowing me to take $5 in the early game, when most players had to spend a pair to take $2), I’m not sure that it mooted the rest of the game, but it felt that way.

Frankly, even as a non-18xx game it was kind of boring. As an “18xx” it was borderline insulting. And why does it have cats on the box? Well, it was sold via the internet. I assume that’s why.

Rating — Indifferent, borderline avoid. Avoid

Update — Lowering rating to Avoid, based on the following anecdote.

The first die roll of the game was 1 1 1 4 5 6, [start player] naturally took the three 1s and used the three cubes on the “1” track tile. This let him connect to three towns from then on he could easily just take any single die on his turn and collect 4 income, while [the other players] could not get nearly that much barring a lucky die roll … and could not even build on our first turn because he had flipped the tile, so now builds were not free.

Yeah, this seems like the sort of thing that a non-ChatGPT game designer should have noticed. “Hey, the first die roll can decide the game.”