Settlers of Catan is a bad game

Settlers of Catan is probably one of the most important board games out there. In part because it's probably the most common "gateway game", i.e. game that gets people into "serious" board games. In part because it's a board game that people who aren't "serious" board game fanatics often happily congregate around. I can recall many people coming to our undergraduate board game club who were only interested in playing Catan, and even just hearing people on the buses making plans to play Catan Friday evening (alongside, you know, the other usual recreational activities of substance that undergraduates do on a Friday night). I too was enamored with Catan at one point during my high school days, and had several consecutive weekends of meeting with friends just to play Catan.

The title of this post tells you where it's going, and I do want to preface the meat of this post by saying that yes, I understand Catan has brought many people joy, that it is fundamentally very important and I and others who have enjoyed the population explosion of board gaming owe much to it. But, there have been many times in my life where would ask to play Catan and I had to turn them down and politely bottle up the many negative feelings about the game that had accumulated during my period of fanaticism for it, so that these people would not think I hated board gaming or that people who played games like Catan were very judgmental.

Thankfully Catan is no longer the popular game of choice amongst my friends, but I still have these feelings bottled up and would like to let them out. So, here's a blog post on the many reasons I think Catan is a bad game and not worth my time to play. This is unfortunately going to be a post that doesn't make much sense if you've never played Catan but... surely you've played Catan, right? I'm also talking about the base game here, the only expansion I've played is the one that adds 2 more players (which only made the game worse, unfortunately).

Let's quickly go over a few reasons why Catan isn't great, and then I'll get to the nail in coffin.

Luck based

You can have the highest expected number of resources earned per round, but no tile generates resources more than 5/36 of the time (which means that no tile generates resources at least once in four rolls with probability greater than 45%, which especially in the early game when you may only have one tile that generates a certain resources is painful). so it's very easy to get screwed by rolls. This isn't helped by the fact that if you lose half your hand if a 7 is rolled while you're over the hand limit, so sometimes you'll get a sequence of rolls that gives you a lot of resources and be punished for it.

This is unrelated to the rest of the post, but I feel like Catan's luck is very feel-bad. It feels really bad to go many turns without getting the one resource you need, and when you get it in a timely manner it's sort of what you're "expecting". If there were a way to make tiles generate resources more consistently and not ruin the pacing I might not mind the luck as much.

Snowbally 

The way to get VP is to build more settlements/cities, which give you more resources to get more settlements/cities, etc. So if the players are playing correctly (see below), the game will probably be mostly decided after one person gets a small lead that can snowball. Sure, the other players can gang up on the leader by blocking their resource gains with the robber but a) you can only block one tile, whereas each player should have settlements/cities next to at least six tiles at any given moment, and b) the leader should have enough resources to hoard knights in defense.

Trading

Trading is a pretty useless mechanic after early-game when played optimally. Usually you don't want to take trade offers once you have enough of an engine going to secure resources on your own. As we'll discuss soon, it's not like you need or want a large diversity of resources in order to win, which is the main thing trading enables, and if you have a monopoly on a resource you definitely don't want to be giving away the advantage of that monopoly. In addition, the person making an offer usually will only make that offer if it provides them a non-marginal benefit. I can remember a game where my opponent was desperately trying to trade for a wheat to finish a settlement while I was sitting on a hand with four wheat. The benefit of not giving him that wheat far outweighs whatever cards they were offering which I probably didn't really need that much.

One way to play

Now, let's establish what is widely considered the best Catan strategy. I'm not a Catan world champion or anything, but I once read a blog post written by one (which I can't seem to find, but if you Google "Catan strategy" you usually find suggestions like this one) so I'm obviously about as qualified. More or less, the key thing to know is that wheat and ore are good and you want as much as you can. Your goal at the start of the game should of course to be to maximize the number of resources you're getting per roll, so you can snowball into the late game. There are two ways to do this, build a new settlement or upgrade a city.

Building a new settlement by adding two roads to reach a new spot and then building a settlement there requires 3 brick, 3 wood, 1 sheep, and 1 wheat. On the other hand, upgrading a settlement to a city requires 2 wheat and 3 ore. Both give you 1 more VP and roughly the same resource gain long-term, but the latter requires many fewer cards, so you probably want to do it as much as you can. So the optimal production cycle in terms of needing the fewest cards is, road->road->settlement->city, which in total costs 3 brick, 3 wood, 1 sheep, 3 wheat, and 3 ore.

This might make it look like you want a balanced spread of resources, but we haven't accounted for many things. First, development cards require wheat/ore and you'll probably want to be picking these up alongside the cycle of building cities to protect your engine (plus, if you can play the most knights, you'll get the 2 VP for largest army, which is nice because conveniently, cities alone can only get you 2 VP from winning). Notably, if you get locked out of wheat/ore by the robber, you're also locked out of moving the robber and are just dependent on the die rolls, so strategies that don't build multiple wheat/ore are more susceptible to getting stalled by the robber. Second, you start with two free settlements, and you should be putting these settlements in the best place available on the map. So your starting settlements are the ones you want to upgrade the most, and at the start of the game upgrading one of these to a city requires less resources/time than trying to build a new settlement. Third, if you end up producing "too much" wheat/ore, you can always trade it in for the other resources. So it's better to take a strategy where you can get a lot of wheat/ore rather than try to have a balanced but much smaller amount of all five resources. I've had multiple games where I was gladly 4-for-1ing away wheat just to get the brick/wood I was missing, because I was so far ahead due to the burst of the wheat/ore start and the fact that my opponents didn't have nearly as much wheat as I did and struggled to do anything but building roads. Fourth, you start with two free roads, so if your opponents don't beat you to the settlement spots you build towards, you can reduce the cycle down to road->settlement->city

So really, the ideal way to win is to build city->city->road->settlement->city->road->settlement->city, which requires 4 brick, 4 wood, 2 sheep, 10 wheat, and 12 ore, and then seal the last two points ideally by winning largest army which will require at least 3 sheep/wheat/ore for a total of 4 brick, 4 wood, 5 sheep, 13 wheat, and 14 ore (regardless of whether you win by largest army or not, you'll want to pick up knights to keep resources coming). You may have to in some edge cases resort to building settlements or getting longest road instead, but because of how snowball-y the game is, once you've slammed down the fourth city, you're probably so far ahead of your opponents that it doesn't really matter. Sometimes you may have to pay a few more brick/wood if your starting roads end up being worthless, but the general mix of resources required is still heavily skewed towards wheat/ore.

Putting it all together

So it's clear now that we need a disproportionate amount of wheat/ore in the ideal path to victory, and wheat/ore also enables the fastest start of upgrading two cities. At this point, you maybe forgot that this was a blog post about why Catan is a bad game instead of a strategy guide. Well, the fact that there is one real strategy to go for compounds with the other factors. This is basically how a game of Catan would go for me once I learned the wheat/ore strategy:

1) Place starting settlements. The quality of these depends on the luck of who gets to place when (and maybe on the tile layout if you randomize that). You might argue that the snake-draft placement of settlements alleviates the luck here, but given the dominance of the wheat-ore strategy, the quality of settlements will probably fall off fairly early in the starting placement process. For example, in the default setup, there are two spots that are really good for the wheat/ore strategy, so the first/second settlements are way better than the rest. Again, wheat/ore are much better than the other resources, so it's not like if you get locked out of a good wheat/ore spot you can take a good brick/wood spot instead and expect to compete.
2) Try to get the best start. Upgrading your first two cities is pretty mindless. Building the next two settlements, it's also usually fairly obvious where you want to go with them (wherever pushes the wheat/ore agenda further) down to the luck of rolls.
3) Once someone has a good start and is generating more resources than everyone else, because of the snowbally nature of the game, they will just win more until the game is over.

As discussed before, trading doesn't really help build alliances in response (usually how multiplayer games handle this) because the wheat/ore strategy is so much better, and doesn't require a diversity of resources, which is all a trading alliance really provides. In addition, there is good reason to be skeptical of trade offers, especially mid-game. So, the first few turns are very RNG-heavy, and then past that the player who has the best resource engine will probably win and the remainder of the game is just an exercise in watching them do so. I'm not saying that I've solved Catan, again, I'm not a world champion. But in this version of the Catan experience:

1) You're not really strategizing too much, you're just using this strategy that people have done the math on and determined is the best
2) Even with the best strategy, you're largely relying on luck-based mechanics to progress it which can be very frustrating.
3) Since the main interactive mechanics are trading (which as we discussed before, is often not worth doing) and the robber (which, once you have the best resource engine, you get to more or less ignore due to knights), you're not really socializing much.

So I stopped playing.

Berkeley Time (and why it, frankly, sucks)

I'm going to try to start making short blog posts on a (approximately?) weekly basis. They'll be about whatever, maybe my work, maybe games I've been playing, maybe just random topics I want to talk about. The main goal is to just to be writing something on a regular basis, so that when it comes to writing important things (i.e. papers) I might be more "in practice". Maybe no one reads them, but that's fine. Most people don't care about people e.g. watching them do practice runs before a marathon, but that practice still has value, so hopefully the same principle could apply here.

Today I'm gonna write about "Berkeley time" (which is not actually a Berkeley-specific phenomenon, but I think it's the most egregious here) and why it upsets me so much. Basically, at Berkeley if a class is scheduled to start at e.g. 1:00, the professor won't actually start lecturing until 1:10. The reason this is the standard (or at least, what I hope is the reason) is because class times at Berkeley all start at XX:00 or XX:30, and class lengths are all multiples of half hours, so it's not uncommon that a student will have a class ending at 2:30 and another starting at 2:30, on opposite sides of campus. The 10 minute delay gives this student time to run across campus and be there for the entirety of both lectures. Again, this isn't unique to Berkeley, I know of other schools that have their own variants.

Before I talk about the easy, obvious fix for this, let me talk about the many many reasons why I find it so annoying.

Unclear what it applies to 
So for lectures, Berkeley time always applies. But, lectures are not the only things happening on a college campus. I attend talks, meet with professors and other students, go to reading groups, go to trainings, go to club meetings, and do many more things. This is as a grad student, the position on the academic totem pole which I think would have the least hard commitments in our schedule (because ideally, most of our time is being spent on research). For undergraduates and professors with much more packed schedules, there are probably many other commitments they have.

Does Berkeley time apply to these other events? I think the subset of these events where it was clarified if we'd start 10 minutes late was a minority, but it wasn't the case that a minority started at the stated time instead of Berkeley time. Without clarification, the time spent on these events is much less useful: If half the people attending something think it starts on Berkeley time and the other half don't, then either the latter half waste 10 minutes waiting for the former, or the former miss out on 10 minutes of this event (both have happened to me).

Inconsistent with other, off-campus events

Now, for on-campus events like the ones I mentioned before, it might be arguable that they should all start 10 minutes late, in case anyone has a class right before these events that would like to attend these events. But, these aren't the only events in our lives. For example, Berkeley has a great resource called the Simons Institute, where researchers working in various areas of CS theory come from all over the world to spend a semester collaborating and sharing their research. There are many great talks that I as a CS theory student would like to attend, but these don't start on Berkeley time (well, usually they don't but sometimes they do and it's not always announced whether they do or not: see previous point) because the majority of people at Simons are not Berkeley-affiliated. So now, if I have class that ends right before a talk at Simons, I'm inevitably going to be late (as opposed to the obvious fix mentioned below, which would give me time to make it to the Simons event).

I also have collaborators from other universities who I'd like to meet with via video chat, and we usually set our meeting times to XX:00 or XX:30 like normal people do. But, this means sometimes I'll be late to these meetings because the Berkeley event schedule assumes it's okay to show up 10 minutes late to an event starting on the hour (or because I'm just wired to show up to everything 10 minutes late: see next point). Oh, and it gets fun when the other campus has its own starting time standard that isn't 10 minutes late (I've seen 3, 5, and 7 mentioned as how many minutes late everything starts on some campuses).

And even for non-work reasons, there are even just events in the town of Berkeley which don't operate on Berkeley time, which means sometimes one needs to rush to them from one's on-campus events or just can't schedule them at a time it seems like one should be able to. Both "important" things like doctor's/dentist's appointments, and more fun things like pub trivia when you need to claim a table.

Sets a bad standard

So before being at Berkeley when I had to show up to things on normal-person time, I usually made it a point to try to be 5 minutes early to events in my schedule, and ashamed to be 1-2 minutes late. Now, I find myself cutting it close more often, and not minding being 5 minutes late so much. Maybe this is just a personal thing, maybe I'm just biased in my perceptions of when I showed up to events before/after coming to Berkeley, but regardless I do feel like the standard of Berkeley time does affect my (and maybe others) tendency to show up to things on time. I also don't think it helps that no one ever officially explained this to me, I just kind of caught on after my first few classes that this was happening. I won't dwell on this one too much, but I don't like that there is an aspect of our campus culture that implicitly is okay with showing up "late" to things.

There's an obvious and easy-to-implement fix

Everything could just start on time and officially end 10 minutes earlier. A 1:00-2:00 event could just run officially from 1:00 to 1:50, instead of officially from 1:00 to 2:00 and unofficially at 1:10 to 2:00.
That's it. An announcement that this is the new standard is literally all that needs to happen to get rid of all the issues with Berkeley time, while allowing people the same amount of time to get between classes and other commitments. The best part is, classes still "officially" start at 2:00 instead of 2:10 in the course registration system, in schedules published online, etc. So it's not like we're officially changing the start time of anything, and if anything updating all publicly listed event times is only being more honest. The only downside is, people have to shift their sleep schedule 10 minutes earlier, but I hardly think that's an issue. It's not like I'm advocating for a revolution of the campus schedule, because under this change it's really not changing at all relative to itself, and it's better aligning with the rest of the world.

I've thought about this a fair amount and can't think of any reason why someone in the administration shouldn't just snap their fingers and make this fairly small change happen (besides, well, the usual gripes people have with campus administrations), and that's really what frustrates me most of all about Berkeley time.