Iterating On Napoleon

The initial test of Napoleon, Blown Apart showed promise, as the game (played two-handed, solo) was actually a fun experience. There were some clear things that needed to be addressed, however. Before I go into those changes, however, it's worth providing a quick description of the structure of the game so you can hopefully follow along with the design decisions. I'll be describing the two-player version of the game here, but it's also designed for three or for four (as a partnership).

At the start of the game, each player is dealt a Reserve, which consists of two Kings for each player plus four more random cards. Eight cards are then dealt as Battle sites, with each card being from 2-8. The game then starts, with the lead player choosing one of the Battle sites. The number becomes the number of tricks in the Battle, and the suit is the trump suit. Each player can then draw as many cards as they want from their Reserve, one at a time, and then the Battle is played out as a standard trick-taking game, with following suit required. If a player takes every trick in the battle, it's a Rout, and the player scores points equal to the size of the Battle. Otherwise, whoever took the most tricks scores a single point. Play until one player wins three Routs or until all Battle sites are done, in which case high score wins.

Now, one of the obvious things that needed to change from this first version is that the scoring didn't quite work. While the reward for a Rout was certainly appropriate, and worked thematically with the overall Napoleonic battle idea, in practice they were difficult to achieve. Three Routs, in particular, was not an especially relevant winning condition. I'm very much a fan of multiple winning conditions, as it can provide a way to enable multiple strategies and potentially provide for surprise endings to the game. But in my first playtests, getting three Routs would almost certainly put a player far enough ahead that they would very likely win on points anyway. The idea was to potentially enable a sudden death win, so that even a player who was behind on points still had a path to victory, but this configuration of scoring wasn't really going to enable that.

So, I changed the scoring to be the difference in tricks taken. So in a 6 trick Battle, a 4-2 split would be worth 2 points, which resulted in a greater difference in scores and some more nuance. I also switched it so that you scored Rout points, and if you got to 7 Rout points, you would win. This was possible in a single Battle, and winning a Rout in a 7 or 8 trick Battle would be akin to shooting the moon in Hearts. But it would provide for hope for a player that's behind, as 7 Rout points was a more achievable target. Between them, these changes gave a more dynamic scoring system, it made each trick matter more, and it provided for more potential drama for come-from-behind victories.

The second problem area with the game is that the Battle selection felt too random. The suit contributed nothing to the decision space, because you had no information about what might be a good choice for you or a bad choice. The size of the battle might be relevant, especially if you already had a Rout, and the size of the remaining reserve for each player might factor in to the decision. But overall, the selection of the next Battle felt very rote. This was intended to be a strategic decision for players, where they could evaluate the current state of the game and make what they think is the best decision for the next fight. That element was just not there, and it was to be a key part of what makes the game interesting.

In order to fix that, the different suits had to be relevant to the decision. In order for that to be true, the player needed to have some piece of private data that pushes them in one direction or another. The easiest way to do that was to provide some kind of preview of their hand. Because the Battles ranged from 2-8 tricks, I could give each player their first two cards in order to fuel the Battle selection. They'd have a limited amount of information to push them in one direction or another. In some ways, this reminded me of a three-handed version of Bridge that I used to play in high school, where six cards of the dummy is exposed before bidding, allowing you to hopefully make an informed enough bid. There were always enough surprises in the remaining seven cards to generate lots of drama, and we ended up calling it "Falling Off Bridge" as a consequence. From that experience, I knew that partial hands would be a fun addition to the game.

Finally, there was an awkward piece of ergonomics in the game. Ergonomics very much do not show up until you try something on the table, which is a valuable side effect of actually getting a game to the prototype stage quickly. The specific thing I was noticing is that when you draw Reserve cards in your hand, you no longer had cards equal to the Battle size. That resulted in a change in how quickly you'd run out of cards in suits, which is a mixed bag for the game, but it could also result in playing the wrong number of tricks if both players drew on their Reserve. On one occasion, I played an extra trick in a Battle, and on another occasion, played an extra card from one hand before catching it. Given that I was paying a ton of attention to how things were going because I was testing, I thought it likely that players would get this wrong sometimes in practice. A fix here is to have players discard a card (face down) when they draw from the Reserve, which also has the nice side effect of letting them better shape their suit distribution, which meant even low ranking Reserve cards might still be useful.

With these fixes in place, I could try out some more tests and find the next layer of things to change. Mechanically, the way that I capture these things is just taking quick notes of problems as I play, without trying to fix them on the spot. It's more valuable to just keep going, taking raw notes of observations as I go, and then consider what I am going to change holistically. That way, fixes can work together. If something is really obviously wrong, sure, it might be worth fixing on the fly. But usually it's better to just complete a game and then consider the changes later.

My goal at this point with the game is to get to a core structure with a standard deck of cards that I'm happy with, one that presents challenging, interesting, and fun decisions to players as they play. I'll focus on the two-player version, as the easiest to test, and then confirm that the rules carry over to the three- and four-player versions just fine. Once I have that core game working well, I'll add some more thematic flavor to the mix. In particular, I want to customize the deck to be more Napoleonic, with suits corresponding to different types of troops, with some notion of commanders and scouting, and potentially some ability to choose your ground for fighting (and potentially modifying the Battle sites). Adding thematic touches is an easier thing if the basics are already in place, because judging the impact of those additions is easier if you know the game is already fun. "That made this less fun" is easy to percieve, and is a reasonable way to judge the chrome you bolt onto a game.

Coming up next time, the results of these tests and the further changes I make, as I continue working on the title.