Gaming the Design - Infamy

Infamy is a game in early design by Travis Chance that centers around an order-fulfillment system (get goods X, Y, and Z and be the first to bring them to the specified location). It does so with a theme that crosses Total Recall with the three-gang system of Grand Theft Auto 2. Player interaction is brought about through both a race to fill the orders (complete the missions) as well as a Pirate's Cove-style simultaneous location selection where overlapping choices are resolved in a priority/bidding system. Accomplishing these missions will gain you special abilities as well as scoring in two parallel point tracks (Infamy and Status). The first player to reach the specified level of either one wins.

One of those point tracks, Status, deserves some examination. The designer had laid that out as three axes leading out from a center (representing the game's three factions) with markers for each level. A player's marker indicated both his or her current Status level as well as with which faction that player was currently aligned.

This layout gave a misleading impression. It made it seem like choosing to support one faction moved you away from others. In reality, anytime you completed a mission for a faction, you immediately switch to that faction without losing any levels. This meant that the factions were almost meaningless in practice. The only thing one cared about in mission selection was which mission you could accomplish, regardless of faction. Factions became irrelevant to your strategy.

To be fair, the designer had a goal of not wanting to penalize players for switching factions, as had been true in previous versions. It is true that players will object to mechanics that penalize them for adapting to the situation. If players were hurt for taking advantage of an opportunity in another faction, they might instead spend a turn or two gathering resources and waiting for an achievable mission in their current faction. This would slow the game down unnecessarily.

As a remedy, I proposed shifting to a bonus-for-loyalty system instead of a penalty-for-switching. When accomplishing a faction's mission, instead of gaining generic Status points, the player would receive tokens specific to that faction. Each token would be worth an increasing number of Status points if within the same faction, but all faction tokens would add up to your overall score. For example, if you had . . .

• 4x Corporation tokens (1 + 2 + 3 + 4 = 10 Status)
• 2x Rebel tokens (1 + 2 = 3 Status)
• 1x Mafia token (1 Status)

Then that would total 14 Status. A player would then be incentivized to be loyal to a faction, but could still work for other factions for a lesser reward. You avoid the psychological effect of penalties while still reaping the thematic and strategic benefits of incentivizing loyalty.

The game is divided into rounds. Each round has the following phases:

1. Bidding for five randomly chosen contact cards
2. Simultaneous-Action-Selection of missions or resource gathering
3. Simultaneous-Location-Selection of missions
4. Resolve Missions
5. Simultaneous-Location-Selection of resource gathering
6. Resolve resource gathering

The bidding system has some quirks. First, you are bidding on contacts that act somewhat like the specialists in Planet Steam. The differences are that you are bidding on a specific one (in their randomly determined initiative order) at a time and that players may win the bid for more than one. Also, you are not bidding with a resource used elsewhere. You are bidding with a fixed number of "bid" tokens that you get back every other round (which means every 10 items). This is coupled with each auction being a "Glum Losers" auction (where all bids, including losing ones, are paid).

This is a dangerous combination. While a Glum Losers auction is both extremely interesting and meets the designer's goal of eliminating frivolous bids, it is an auction type for which most people are unprepared and that severely punishes those unfamiliar with its results. A famous example of this is a professor who auctions a $20 bill to his students through this auction type where individual bids eventually go well beyond the value of the prize. While this seems impossible, it is the predictable result of rational people trapped by this insidious action type.

You can skip this part if you are already familiar with the story of how this happens. Michael bids $10 for the $20 bill, hoping to gain a $10 profit. Jeff then bids $19, knowing he will reap $1 profit by winning and assuming that no one will bother bidding more when they cannot net gain anything by doing so. However, Michael is already in the trap. Because of the Glum Loser rules, by losing the auction, he will end up down $10. Michael increases his bid to $20, because this will mean that instead of losing money, he breaks even. Jeff is now in the trap as well. His now-losing bid of $19 means that is is perfectly rational for him to bid $21 for the $20 bill. While that would mean he has a net loss of $1, that is $18 better than losing his previous $19 bid and getting nothing in return. Both players are now in a cycle where it is perfectly rational for each to come over the top as while both players are now guaranteed to lose money, neither wants to be the one to lose more money by bidding less.

The danger of the Glum Losers auction is even worse when you take into account the way bidding resources work (fixed number per player, regenerated every 2 rounds). If two or more players get caught in the Glum Loser trap early, they will all end up with nearly nothing to bid with for up to nine more auctions.

The designer felt that previous tests with this system had gone well. I felt that he was discounting the danger of a new player digging themselves into a deep hole right at the start of the game. Such an experience might turn that player off the game for good. I personally like alternate auction formats (the keep-goin'-round-forever bidding of Power Grid is the low point of one of my favorite games) but am acutely aware of how important first impressions are to a game.

The prediction of doom bore out. In the third auction two of the four players got into a bidding war and were essentially eliminated from bidding for the rest of that round and all of the next one. I was able to manipulate that into winning five of the ten auctions between the first two rounds and the opponent who had not been caught claiming three of them. These won auctions gave a massive lead in resources that could be used to complete missions.

The designer did not want to switch to normal bidding, because that would go against his goal of eliminating frivolous bidding. I suggested a switch to the Penny-Auction model. In this model, losing bids are returned, but merely placing a bid incurs a nonredeemable cost. This auction type ensured that:

• Frivolous bids were still eliminated, because merely placing them had a cost.
• Players were encouraged to bid their maximum value for the item immediately, as having to increase the bid after being overbid meant spending yet more just to bid a second time.
• Since losing bids (minus the per-bid-cost) were returned, the rationality trap of Glum Losers would disappear.
• Even if a player fell into a bidding war, he or she would lose at most half of his or her money instead of all of it. Only the price-per-bid was lost, the losing bid itself would be returned.

Another goal of the designer was to have a high rate of collision between players during the mission/gathering resources phases. One method of doing this was by hoping players would simultaneously select identical locations in an attempt to both get the resource (or mission) provided by (or completed at) that location. Two things got in the way of this goal.

The initial obstacle was that the selection of location was broken into three pieces. First, players simultaneously selected whether they wanted to accomplish missions or gather resources. Then, only players who selected missions would then simultaneously select locations and then resolve conflicts. Finally, only players who selected gathering would simultaneously select locations. This meant that you would only potentially be in conflict with about half of your opponents (on average).

The other obstacle to the goal was the sheer number of locations. The game started with 14 locations with the possibility of more being built. This meant that even if you were potentially in conflict with other players, it would still be relatively unlikely for you to collide with one of them.

Collisions are resolved by which player had won the auction for the highest-initiative item. However, the player who would lose that comparison may discard an ammo item to try to win the location. The other player may then do the same to cancel it out. Since effectively losing a round's action by failing to win a collision was damaging, players would be incentivized to go back and forth with this.

The problem with this resolution system was that since resources were public, both players could see who would win the conflict by how much ammo each side would have. Therefore, the player who could not guarantee victory should logically not fight at all, since that would just make them lose ammo without gaining anything. An argument was made that a player might do so anyway out of spite, but this is clearly not designed to be a take-that game. It is designed to be strategic, so it should be judged on the soundness of its mechanics subjected to strategy-focused players.

There is another argument for spending ammo when losing, but it was a marginal case that would happen in the rare event that a collision occurred during mission resolution with an ammo-requiring mission on the table that referenced that conflict's location.

Thus the location collision mechanism was both rare and, even when it did happen, strategically uninteresting.

I proposed an alternative system where, instead of simultaneous-action-selection, players would select a location in initiative order. You could only select a location already occupied if you spent an ammo to kick that player out. A player who was kicked out could then either choose a new unoccupied spot, or take an occupied spot (including his or her former one) by spending an ammo himself or herself. The benefits of this change would be:

• The value of ammo is identical to the old system.
• The value of initiative is identical to the old system.
• Collisions are no longer random but through tactical decision making.
• Collisions will be exactly as frequent as players want them to be.
• Even if you have enough ammo to win back a spot, it might be advantageous to switch to an empty spot instead of fighting back because you no longer lose a full round's action by forfeiting.
• Because of the previous point, it is tempting to initiate a fight even if you have less ammo, because the other player might not fight back.

While I liked this suggestion, I realized that the designer might not want to ditch the simultaneous-action-selection. So I suggest that, no matter what, the total number of locations had to be reduced to increase the frequency of collisions.

The designer's other collision mechanism was that each mission card had three missions, one for each faction. This meant that players could be going for two different parts of the same mission card, but whoever had higher initiative would get to do that third of the mission card and the other two thirds would be unavailable.

While this is a good way to accomplish the collision goal, it has the downside of making mission cards full of 15 pieces of information: Status reward, Infamy reward, special effect, and faction icon / location / required goods / faction reward for each of the three factions. Multiply this by three mission cards and there was so much information available that players were overwhelmed.

I ended up suggesting that the designer only have one mission per card to deal with the information overload. However, this would go against the designer's collision goal, so this suggestion was a failure. I still don't know how to square this circle of information overload versus encouraging collisions.

The takeaways from this game are:

• If players don't like a penalty, a good alternative is to reward the opposite behavior.
• Odd auction types make a game interesting, but they also risk players not understanding the subtle strategic differences between auction types.
• If you want players to interact, try to maximize their ability to choose to do so. On the flip side, minimize their chances to avoid interaction.
• Think about how a perfectly rational player will approach a mechanic.
• For playtesters: point out every problem you see, but when making a suggestion to fix it, keep the designer's own goals in mind. Help him or her make the game he or she wants to make.

That'll wrap up this edition of Gaming the Design. Do you have any questions about this game? Post them in the comments and the designer or I will attempt to answer them. The designer has already sent a response, but I'll put that in a separate post tomorrow as this is a tad long already.

Would you like to have your game featured in a future Gaming the Design column? Let me know at gdw@gamedesignerwannabe.com