analytics code

Wednesday, May 16, 2012

drafting with hidden attributes

As a follow up to my previous entry, I want to consider some more specific themes that could be used with a drafting mechanic that utilizes hidden information. The drafting mechanic discussed here is not the typical "pick one and pass" card drafting that other games utilize. Instead, it's something more commonly seen in both real and fantasy sports. There is a known pool of items to draft from, and players take turns picking an item to add to their set of items.

If the draft is the meat of the game, the majority of the time spent playing the game should be spent during the draft. Resolving the draft -- i.e, determining which set of items is the best -- should be a relatively quick exercise. Ideally, multiple drafts would occur, and the winner of the game would be based on the outcomes of each competition that occurs between drafts.

The most obvious theme is drafting and managing a sports team. The obvious benefit to using this theme is that it closely mimics what happens in real life, and there are millions of people that play fantasy sports every year. The problem with sports is that most sports require fairly large teams. Of the major U.S. sports, basketball requires the least number of players with only 5 playing at a given time. Even then, 5 players for each team with 3 - 8 teams requires 15 - 40 players available for each draft. That's a lot of players to consider at once!

Sports do play well into the theme for other reasons. Players generally specialize in a position or skill that every team needs, and each team would need to have players that fill all of the needs of the team. Players can have synergy with other types of players that would enhance their value to the team. Using dice, players can be represented as providing consistent value, or they can be inconsistent but have high potential -- adding a bit of risk-reward to the situation.

Finally, the hidden attributes: each player could come with a randomly assigned card that slightly modifies their attributes, but the contents of the card are unknown to the teams. A team can hire a scout to allow them to look at the contents of the cards for a particular player. Teams with more money can hire more scouts. The teams that perform the best between drafts (in either head-to-head matches against other teams, or perhaps in some calculated ranking) earn more money to use for scouting in the next draft.

I could expand further on ideas I've had for doing a sports drafting game, but I want to focus on one other route that I've considered.

Instead of building a team, players are building a machine. Each machine requires certain parts to run, and each part affects the performance of the machine. Each machine also has a maintenance cost (again, randomly assigned) that is hidden from the player, but can be revealed to a player if that player sends one of his "experts" to examine the part. Each part has an initial cost that the player would have to pay when they pick the part in the draft. Before the next draft, they would have to pay the maintenance cost or trash the part. The game would flow something like this:

  • Create pool of parts for drafting
  • Send out experts to look at hidden maintenance costs
  • Draft parts, pay cost, reveal maintenance costs
  • Compete, earn payout based on performance
  • Create new pool of parts for drafting
  • Pay maintenance cost or trash each item (not sure when exactly this should happen)
  • Send out experts, etc.
So what are these machines?

One idea is to make it an auto racing game. I know nothing about racing, and I haven't played any board games based on the theme. However, the advantage of a race is that the results are determine in one race! You don't have to have several head-to-head matches. No matter how simple the method used to determine the winner in head-to-head matches, that process will almost certainly take longer than running one race. Additionally, defining how the head-to-head matches should take place between drafts can be difficult with variable player numbers, unless you do a round-robin. And nobody wants to do an 8-player round-robin.

The other idea is robot combat. Yeah, you heard me. ROBOT. COMBAT. Now the head-to-head matches don't seem like such a bad idea, do they? An interesting aspect of this theme is that players could design their robots with a particular opponent in mind if they were only facing one or two other robots between each draft. The makeup of your opponent's robot becomes one more factor in the decision on which parts to draft. 

Perhaps it should be a Robot Battle Royal.

So what sounds good? Sports? Auto Racing? Robot Combat? Robot Battle Royal?

Friday, May 11, 2012

distilling the nfl draft

A recent tweet reminded me that I have a "fantasy sports as a board game" design that I was once working on.

Let's break this down.

The purpose of the NFL draft is to give each team the opportunity to add pieces to their team in order to improve it.

The draft adds pieces to a whole with the intent to improve it.


The order of the NFL draft is determined by the performance (win-loss record and playoff performance) of the teams in the previous year. The teams with the worst performance select first. This is necessary as the draft is inherently unfair. The first selection is more valuable than the last player.

A draft is inherently unfair, so the drafters must start in unequal positions.


The players drafted are evaluated on a variety of aspects. They have shown their football abilities at the college level. They have shown their athletic prowess at the combine. They have a perceived potential. The position that they play can add to their value. Each player has a better fit in some systems than they do in other systems. The player's character is evaluated as well. All of these are factored into where a team will rank the player on their "big board".

The pieces have a known value, a potential value, and a value that is dependent on the whole they are being added to.


There is no winner after the draft is completed. Instead, the results of the draft are determined in a 16-game season, followed by a 4-round playoff. While favorites can be identified based on the make-up of the teams, there is enough unpredictability that the champion cannot be determined until the entire season is played.

The results of the draft do not determine the winner. Instead, each whole is has the potential to be the winner, but some are more likely than others.


But what about the "subterfuge and meta-gaming"?

Each team has its needs and has its own valuation of the players available in the draft. If there is a player that a particular team wants, but they are afraid that the player won't be available when it's their turn to pick, they will make a trade -- players or future draft picks -- to move up to secure the player that they desire. So when another team feigns interest in a player, they can convince a team to make a trade, perhaps unnecessarily, to acquire the player they want.

Can this be built into a game? Or does a game rely on players to create the meta-game?

A lot of this relies on the players. The design can encourage bluffing and posturing, but it can never force a player to act in such a way. Werewolf, Mafia, and games of that variety are built entirely around meta-gaming. If a player refuses to lie or otherwise trick and deceive the other players, the game will be much less interesting.

Auction games afford some amount of meta-gaming because players have the ability to affect the price for items that other players want. A player can force another player to pay more for an item that is only valuable to one of them. In the same right, a player could be caught bidding another player up too high, and be forced to pay more for something that isn't as valuable for that player. Games like Homesteaders and Power Grid excel in creating that tension of one player seeing how far they can push up the price for another player. Ra not only forces you to judge how much a player will bid on a set of items, but it also allows you to choose when to offer some items up for auction. In the right group, extensive meta-gaming can emerge from these simple mechanics.

Coming back to the draft, I believe that while meta-gaming could emerge from the uncertainty of how each player values a particular draft pick, the addition of hidden information would encourage meta-gaming even more. Imagine a draft where each item in the draft (e.g., a player in a football draft) has a set of know attributes that all players can use to judge the value of that item. However, each item also has some hidden attributes that not every player has seen. Players that have this information can use it to get a more accurate valuation of the item, and they can deceive other players in what they saw. If more than one player has seen the same hidden information, the two players can create confusion if one is trying to tell the truth about what they saw while the other is deceiving the other players. While this still requires the players to create the meta-game, it adds another layer to the design to encourage it.

Since this post is already too long, I'll save my game ideas for drafting games for another post -- hopefully soon!