Notes from New Bedford: Part 16 – Removing a Player, Adding a Captain

Index

The solo mode was the last thing to be added into New Bedford. But it was not a last minute addition. I had already started considering a solo mode before the initial campaign launched. At the time it was rough, and the extra few months I got before the relaunch enabled me to develop it considerably.

When I initially considered a solo mode, I knew it would not simply be a race to points. I read a lot about solo games of Agricola. But unlike Agricola, there aren’t hundreds of cards to vary the setup to make new puzzles. New Bedford specifically tried to minimize that aspect without giving up the variability. With only real output randomness in the game coming from whale draws, it would be simple to calculate an optimal building score. In fact, I had done many calculations while checking the balance.

New Bedford’s variability comes from timing. Not only when players use buildings, but when and what they build. So any New Bedford solo play would have to add buildings at different times, and be able to adapt to the buildings in use, because blocking action spaces was a key element of strategy. I was going to need some kind of AI that felt like it knew what it was doing. So arranged the basic actions in a rondel. Sort of a logical pattern, so that you’d move ahead a space and if you couldn’t do the action, you’d move ahead again and try something else.

This was way too predictable. You could easily win by using the AI’s next action one turn ahead. Having the AI always go first helps this, and that rule would be added later. But what I really needed was some way to randomize it. But that logical order was appealing, because it reflected actual action patterns. Collect wood, prepare a ship, collect food, launch. You can usually guess what a player will do, pure randomness wouldn’t capture this. So I added a biased die. The AI usually took the next action, but occasionally jumped ahead.

This worked reasonably well, most of the time. The AI could build based on available resources and simple instructions, which varied what buildings came into play. But sometimes the AI would simply get stuck taking a stupid action, like collecting wood for the entire game, forgetting to launch a ship, collect money to pay for whales, or ignoring a bonus building. It seemed like tracking the AI’s money was problematic and didn’t add much, so I tried having the AI cheat a little by ignoring money.

After a short time, I discovered that the real bookkeeping problem turned out to be resources. My solution was to have the AI cheat more, and track only money instead, which worked reasonably well. The AI collected a little money instead of resources, which felt like it balanced out occasional poor decisions in what buildings to use. The AI always had enough resources to build, but that meant that I needed to tell players which buildings get built. Having two lists meant that you couldn’t rely on the AI building a given building in a game.

The game stayed that way for a while, but leading up to the relaunch, experienced solo gamer and designer Mike Mullins really helped me to refine and balance the solo mode much more. Not being a real solo game player myself, he showed me a lot of what I needed to change to make it fun for a solo player.

Up to this point, the AI simply paid for whales immediately, which occasionally changed which whale it chose and reduced bookkeeping when the AI’s ships returned. But Mike wisely pointed out, that didn’t really feel like the full game as much. It wasn’t much extra trouble to track whales, and since the AI didn’t change its action based on how much money it needed, it wasn’t any extra bookkeeping, and made several buildings easier to use.

Mike also pointed out the problem with building lists that I hadn’t noticed in my few plays. On average, the AI would build the same buildings in the same order. Even though the order made sense, it made the game less replayable. Mike suggested randomizing all the buildings in play and having the AI build them in that order, but that felt too random. I eventually took that idea and combined it with the building lists, so the AI builds things that work together, but it will change from game to game.

In testing, the AI basically always had enough money to pay for whales, use buildings, and had plenty left over at the end of the game, even when they started with nothing. The AI was getting a double dose. Not only did it have enough to pay for whales and build everything, it got a bunch of points for doing nothing. In short, the AI cheated a little too much. So one of the last changes was to remove money completely.

The part I haven’t discussed yet is the Captains: I also knew early on that I could make multiple rondels, with different sets of icons to emphasize toward the three main strategies: balanced, building, and whaling. I considered the possibility of wanting to use them together, so by using different sizes for the rondels, they wouldn’t line up repeatedly. This is where the Captain personalities started, with Starbuck, Flask and Stubb.

I also needed to guide their actions, and added some simple instructions for how to pick a building to use. Different captains prefer using different buildings that support the general strategy provided by the rondel. They also include different lists of buildings to build, which further reinforces their strategy. I added some unique instructions about launching and what they try to do in the final round, because they would frequently ignore bonus buildings that are left.

Even after removing money, the captains still needed to be limited a bit. To make them less powerful would result in frequent wins, so I looked back at the hundreds of games I played and found some upper limits for what was possible for a player to achieve. A high scoring balanced player would frequently top out at less than 10 whales and 6 buildings. An expert whaling strategy (and the captains are certainly experts) could potentially pull in up to 15 whales, but at the cost of buildings. A master builder can build about 10 buildings during the game, but that number is reduced by trying to pay for whales late in the game.

I could fit the rondel and instructions on the back of the player aids, making it a great addition to the game. But three captains and four players felt like someone was missing. Enter Ahab. While the other captains took a well-reasoned approach with logic to their buildings, Ahab was the wild captain with a balanced but unpredictable strategy, and more powerful, too. Rather than following a set course, if he is unable to find the building he wants, he moves again. If he is unable to find a second building, he skips his action and draws a whale from the bag. And of course, he always makes sure he has ships at sea in the final round. And, of course, if playing with the White Whale, Ahab always gets it.

Because the captains simulate players to a large extent, it raises two possibilities. First, adding multiple captains to a solo game. And second, to add a captain to a two- or three-player game, to make it larger. In my 2-player testing, we had a great time yelling at the captain, who always seemed to take the exact action the two of us wanted. My hope is that these Captains’ personalities shine through, and you can have as much fun yelling at them as we have.

 

Advertisements

,

  1. Lessons from Designing a Solo Variant | Oakleaf Games

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: