Random thought about unit characteristics

In table top gaming we are living in a culture of numbers and calculations. Almost any wargame or miniatures deals with its subject matter by some sort of “numbers comparing”.

Units are arrayed on a 2 dimensional plane, which has a visible or invisible grid of coordinates. These units have characteristics expressed in numbers like morale, ability to inflict damage, movement distance and so on and so forth. The player changes the game state by announcing an action and feeding the numbers through an engine to get a result.

Maybe it is time to break this paradigm of number crunching by swapping it with a narrative description engine. Well this is just a hazy idea at the moment, but when reading the beta rules of the Drowned Earth I immediatly questioned my gaming efforts in the last months.

Continue reading “Random thought about unit characteristics”

Advertisements

Analysis of second test game

I just posted a rough outline of what happened in the second test game right here. All in all it was satisfactory from a system standpoint. Let’s reflect on three things that …

… worked:

  • I finished the game in 1 hour. One week of campaigning in northern France. At the end of the game/hour supply would’ve been checked, attrition and new troops calculated and media/politics effect randomised and experience allocated. After that a new game woukd’ve begun other on the same board – fe Austrian counter attack- or another board – fe adfter their success french commander s ordered to cross the rhine.
  • The bot moved well even on the attack. Generally I feel like defending is easier for the hot than attacking. Moving a decisive force to a specific point to “win” actively is harder to simulate than the reactive movement of moving forces to the defensive position in danger. Overall it kinds worked. The Austrians could’ve attacked Nancy in force, which probably would’ve been enough for an Austrian win. That they did decide to change orders for their leading column in the last moment was brilliant and rather Austrian at the same time.
  • The asynchronous and non linear passage of time for the bot produced enough operational friction for me to feel like i was in command. As a player I knew roughly how far my columns move in one phase. I am never sure whether my scouts and subordinate commanders tell me what’s happening though. A3 moved extremely fast through the middle of the map – a surprise move – a triple activation. In narrative terms the french scouting cavalry went to the inn instead of doing the job thus the commanding officer of the french only got the message of the approaching Austrian column when it was right at the door step of Nancy, leaving almost no time to react. This relativ Passage of time for different columns, combined with the fact that the limited resource is one’s ability to change orders really emphasises the huge delays in the ooda loop experienced at the very top of 18th century armies.

What didn’t work:

  • Combat takes way too much time. Multiple steps and a lot of dice rolling just isn’t for me. I want to know the outcome and that’s it – or is it? Still not sure about this. Combat in the age of Napoleon was not a linear affair. The bigger force didn’t win and inflict x amount of casualties 90% of the time …. Different battle tactics and actions could change the outcome dramatically. Now I don’t want to simulate this as my focus is operationally on the other hand I need a system that can produce a wide array of possible outcomes.
  • Morale system should be further integrated with what columns do. The threat to the Austrian supply line should’ve lowered the morale, making them switch to defensive orders. The Austrians especially were renowned for over protecting their supply lines.
  • Some parts of the rules are still too wishy washy leaving room for me to interpret them creatively. Especially the beginning and end process for the “campaign” needs to be fleshed out.

What i still didn’t do:

  • The characters still do not have names, characteristics and qualities that carry over from game to game. Before my next test game I need to design me some characters to lead the french.
  • Supply still is an afterthought, as i didn’t play long enough for it to become an issue – basically at the end of a turn a number of supply checks is to be done but i did only play one turn
  • Divisions are just numbers, they too should be infused with some character.

I am positive that my next test game will be game 1 of my campaign. The campaign will start in early 1792 and will cover the rise and fall – depending on my luck – of a young fictional general. Each play will deal with one campaign season in a distinct theatre. Depending of the success or failure of the young officer the theatres will either become bigger and more meaningful to the overall fate of france or he will be relegated to obscure and unimportant parts of Europe. Let’s wait and see. Oh and the notebook has tags now. Hurray!

Thoughts about bots.

Solo games can have completely different systems for players and the bot. In effect the bot needs to do things that are challenging to the players, that cannot be gamed and that can be interpreted as being realistic.

So while as a player I like the challenge of assigning ever too few operations to my columns, the agony whether I want to press my attack on the left flank but leave my right flank woefully unprepared or whether I want a more balanced approach, the bot does not need to care about these things at all. The bot just does things that have effect.

To put it short and simple: the bot produces results instead of simulating a mind that takes choices. If the bot is in fact not playing the game but just a way to build the game for the player most of the challenges with “making the bot play well” disappear.

Instead the problem of making the bot “feel realistic and fair” take the forefront. My experience tells me that most players have more difficulty accepting the actions of a bot if the bot does not seem to play the same game. In the coinseries, for example, bots play pretty much the same game as the player – controlled by an elaborate decision making flowchart. I tend to interpret these flowchart-bots as challenging and “fair”. There is after all no point in the flowchart which says:

3 divisions magically teleport to Milan*

But what I can do though is deconstruct the way the bot works. Analyze its flaws and strongsuits and “game the system”. The bots in the coinseries are clearly built with a lot of care and thus are not easy to game, but for me as a player the experience shifts. The goal for me as a player moves from “winning the game” to “learning how to game the bot”. Not something I want to fill my little gaming time with.

1944 race to the Rhine on the other hand works quite differently. The appearance of German troops follows a system uniquely for this specific task. This system is random and behaves somewhat erratic. I feel a lot less inclination to learn the subtleties of this bots behavior. But, and this is a big but, I tend to interpret the results as unfair, because they do not conform to the same rules that I as a player conform to, but they are the outcome of a distinct and recognisable subsystem:

What would have happened if I had drawn a different card right now. Oh my I surely am unlucky today.

The challenge for me is thus to create a bot that is interwoven so deeply into the games core systems that I as the player cannot interpret it’s actions because the

  1. underlying logic is not apparent
  2. There is in part no logic at all

… gotta go. I wonder whether this will become a blog of half finished posts 😁

*All my thoughts are still firmly set on the Italian theatre of war even though I am currently playing the defense of northern France in 1814.