Series Info...Trials, Triumphs & Trivialities #144:

Strategic Insights: Auction Games

by Shannon Appelcline

In the halcyon days of early 2003 I wrote an extended series of articles in TT&T about strategic games: in other words, games more like our own Space Federation and GE: Hegemony than our more typical roleplaying games. Much to my surprise I found a real kinship to the strategy games that I was somewhat idly playing at the time, and thus they've become a much more regular part of my gaming life in the year since.

In recent months, one segment of strategy games has particularly caught my attention: auction games. This is in large part due to our own Queen's Necklace and Fist of Dragonstones, each of which approach auctions from very different directions, but it's also due to the fact that some of the most lauded German games continues to be auction-dominant, from a wide array of wide-respected designers.

Thus, I've decided to give auction mechanisms in strategy games a more complete look. I have touched upon this topic before, both in Trials, Triumphs & Trivialities #116, Strategic Insights: Trading Games and Trials, Triumphs & Trivialities #137, Social Gaming Interactions, Part Two: Competition, but each was quite abbreviated compared to what I plan to write today.

Defining Auctions

Contrary to what I might have written in TT&T #116, auctions actually aren't a subset of trade. Rather, they're a closely related but very distinct type of competitive social interaction.

Overall, I'd define auctions as having the following characteristics:

  1. Usually involve players bidding against each other to buy a resource from the game system (not from another player).
  2. Usually involve an exchange of a monetary marker for a resource, or else (more rarely) of a resource for a monetary marker. These are not barters or exchanges of like goods.
  3. Usually involve rigid methods for making bids and changing prices.
  4. Usually involve a price either increasing or decreasing over time.
  5. Always involve the best price being taken following the end of the auction period.

If you compare this to trade games, you see that the typical trade game doesn't meet any of these criteria because: it's more often players trading with each other, not the system; trades are more often barters of like resources; trade methods are more often freeform; and the trade price doesn't necessarily go up or down because the player isn't actually required to take the best offer (and, in fact, since trades are more often barters, there may be no way to define a "best" offer).

Prior to writing out what exactly an auction was, I had some problems figuring out the line for trading games with rigid rules such as Reiner Knizia's Res Publica. In that game there's a very auction-like mechanism where a player offers up a card and takes one offer from each player. However, it only actually matches my criteria #3 for auctions, not the rest. It's player-centric, it's a like exchange, and the player takes whichever offer he wants.

Real auction games will probably match 3-4 of these criteria, minimum.

An Auction Done Simply

The most simple form of an auction, as typically represented in a game, goes something like this:

An item is offered up for sale and each player, in turn, either makes a higher offer than the previous player, or else passes. Auction bidding continues going around the table until all the players but one have passed, and then the item is sold to the last bidder for his bid price. The rest of the players do not spend their money.

This very simple form of auction can be found as the basis of Tom Lehmann's year-one auctions in Pizarro & Co. and Reiner Knizia's High Society. These games vary as to whether a player who has passed can get back into the auction or not, but beyond that they follow the core mechanism pretty well.

Some of these simplistic auctions work fine, but games can also be a lot more interesting if they use auction methods that are more complex. In the rest of this article, I've outlined a number of different ways to classify auction methods. They include: bid timing methods, bid display methods, bid valuation methods, bid resolution methods, and bid limitations. I'm then going to finish up the article with some advice on how to make a good auction game and some examples.

Bid Timing Methods

The first question in any auction game is: how do you organize bids?

Freeform Bidding: Bidding is totally freeform, with anyone able to offer a bid at any time, and bidding continues until no one has offered a bid in sufficient time. This is how "real" auctions tend to work, but is rare in auction games. Reiner Knizia's Modern Art has four types of auctions, determined by icons on individual cards. Its "cross" auction allows freeform bidding and it works because a player actually acts as the auctioneer, and thus can try and drive up prices and determine when bidding is "done". Another game with freeform bidding is Reiner Knizia's Merchants of Amsterdam and that works because it's a Dutch auction game with a clock counting down until someone takes the price by punching the clock (or else gets their hand broken by someone else trying to simultaneously punch said clock).

Simultaneous One-Time Bidding: Everyone gets to make only one bid, and all the bids are revealed simultaneously. This goes hand in hand with Blind Bidding (see Bid Display Methods, below) and is the core of a couple of closed-fist bidding games such as Corne van Moorsel's ZooSim as well as the "hexagon" bidding in Reiner Knizia's Modern Art. In cases of ties, some other method is used to determine the winner (for ZooSim, ties are alternatively won by each player while in Modern Art it depends on seating position from the auctioneer).

Simultaneous Multiple Bidding: As with Simultaneous One-Time Bidding, but multiple bidding rounds are possible if the first one ends in a tie. In Bruno Faidutti's Fist of Dragonstones a second bid is taken among the tied players using special silver pieces rather than gold (meaning, effectively that each player gets to win at least one tie, because the silver are limited resources). In Martin Wallace's Und Tschuss a second bidding round is done among everyone, meaning past low bidders still have a chance to win.

Turn-based One-Time Bidding: Everyone gets make only one bid, but bidding is done one at a time, usually once around the table. This offers some interesting dynamics because players early in the bidding process have limited information while players late in the bidding process have complete information. Who bids first has to rotate for this to be a "fair" method. This is the core of Reiner Knizia's Medici and the "circular arrow" auctions in Reiner Knizia's Modern Art It's also the heart of less-obvious auction mechanisms, such as the turn-ordering auctions for Alan R. Moon & Aaron Weissblum's New England and Wolfgang Kramer & Richard Ulrich's El Grande.

Turn-based Continuous Bidding: Everyone gets to continue bidding around the table, one at a time. After passing bidders may either be out for good, or able to continue bidding on later times around. Bidding continues until everyone but one player has passed. This is my least favorite auction mechanism for strategy games because it's so unbounded. Auctions can start low and go on and on and on. Some of the aforementioned "simple" auctions use this method, such as Reiner Knizia's High Society and Tom Lehmann's Pizarro & Co..

Bid Display Methods

The question here is: how are bids displayed?

Open Bids: All bids are seen by all players during the bidding process. Most games work this way. Blind Bids: Bids are all revealed simultaneously, which means that the Bid Timing Method almost has to be Simultaneous One-Time Bidding, as discussed above.

There are other possibilities that I've never seen used, but could be:

Partial Blind Bids: Only some aspect of the bid is known to other players (e.g., how much "gold" was bid but not "silver"; how many coins were bid but not what type; what the total value of the bid was, but not its composition, etc).

Totally Blind Bids: An impartial arbiter looks over all blind bids and then awards a winner, but the bidders never know how much their opponents bid, just who won.

I'm not sure if any of these latter methods have any "fun factor" in them; I merely record them to more accurately display the whole continuum from all information to none.

Bid Valuation Methods

The question here is: how do bid valuations change over the course of an auction?

Unconstrained English Auctions: In this type of auction, bids slowly increase, but at a speed set entirely by the bidders. Players may make minimalistic increases or else may make large jump bids to try and quickly end an auction and stun opponents into submission. Again, this is the most common method for auctions.

Constrained English Auctions: In this type of auction, bids must instead fit certain criteria. Perhaps bid increases must always be in certain minimum or regular units, and perhaps those minimum units increase the higher the bid is. Reiner Knizia's Amun-Re uses this sort of constraint to good effect. Bids follow a simple formula: 0, 1, 3, 6, 10, 15, 21, 28, 36. You can jump-bid if you wish, but the constraints mark minimum increases that you must meet at each level.

Constrained Dutch Auctions: Here, no one is bidding. Rather, an item has a price which is slowly decreased until someone decides to purchase it. The first bidder who does so is the "winner". These auctions are constrained because price drops tend to follow a pre-set formula over a pre-set time period. Reiner Knizia's Merchants of Amsterdam does this with a clock which ticks down prices while Bruno Faidutti's Queen's Necklace reduces the value of each item for each player-turn that passes as the auction proceeds.

Variant Dutch Auctions: Instead of decreasing the price, Dutch auctions could instead increase the value of the item until someone purchases it for a set price. I'm not sure that "Dutch Auction" is really the proper term for this, but it's so close that I've used it anyway. Andreas Seyfarth's Puerto Rico does this with its "player role auction". A number of roles are put out for selection each turn, and whichever ones aren't taken at the end of the turn get gold added to them to increase their value. I haven't seen this method in more standard auction games, but it should work just as well as any standard/constrained Dutch Auctions do.

Set-Value Auctions: The auctioneer sets a single price for an item that does not change and the first person to agree to that price wins the item. This is the case in Reiner Knizia's Modern Art "dollar" auctions (where people bid in turn around the table, to take the item or not).

Bid Resolution Methods

This is a combination of questions: who wins the auction and who pays?

We start off with the question of who pays:

First Payer: Most commonly, only the bidders who actually "won" have to pay. This might be a single winner or multiple identical winners. It's how real auctions work--and most auction games too.

All Payer: Everyone has to pay what they bid. Clearly, this sort of thing would never fly at a "real" auction, but is a perfectly good mechanism for a game where you want bluffing to be a serious component. Sometimes "losing" bidders get some reduced reward, while sometimes they get nothing. In Bruno Faidutti's Fist of Dragonstones and Bruno Faidutti's Queen's Necklace jewel sale, additional bids are simply lost. In Reiner Knizia's Amun-Re favor auction, the first-place winner gets the best result, the second the second best, and everyone else a marginal result.

For auctions where there is only one winning result, there is only one choice to see who wins:

Single Winner: Whomever bids the highest wins.

However, there are also possibilties for multiple winners. The first three possibilities cover the situation where multiple, identical items are available for purchase (or sale):

Multiple Identical Winners: Some (typically simultaneous) auctions may allow multiple winners if they all bid the same price. In this case they all get the resources for their bid (or all earn the money for their sale). This method can be found in the price-setting phase of Christophe Boelinger's Fantasy Business.

Multiple Lowest-Cost Winners: Bids are taken via a normal method--most frequently simultaneous and blind--and then all the items are sold at the lowest successful price (e.g., the lowest price for someone who should have gotten an item, based on quanity). For example, if there were five bids for an item, $2, $5, $7, $8, and $15, and there were two of the item available, both the top and second bidder would get the item at $8. This is the auction method type that eBay erroneously calls Dutch Auctions. (It ain't.) It's also very similar to the "Vickrey Auction", but there the highest losing bid is used as the price, rather than the lowest winning bid.

Multiple Discriminatory Winners: Exactly as the Multiple Lowest-Cost Winners, but each winner pays what they bid until supply runs out rather than the lowest price. I'm not aware of this being used in games but it is often used for refinancing credit in the financial world, despite issues of collusion.

The last winner possibility relates to a case where multiple, similar but not identical items are available for sale. As with the case of "all payers" (which it usually goes hand-in-hand with), this is the type of thing that only tends to work in games:

Multiple Variable Winners: You might not have multiple identical goods, but rather have multiple similar goods, where some are clearly better than others. Here, you can have multiple winners who each gain a different value of item, depending on their bid value, with the best bidder getting the best item, the second best the second, etc. Many of the "odder" auction methods have results like this. In Moon & Weissblum's New England your bid determines your turn order. In Reiner Knizia's Amun-Re, your bid in the favor auction determines how many rewards you get, from 1-3.

Finally, there's possibilities for other outcomes than simple winning:

Discrete Results: Some auctions may have discrete results, somewhat separate from the auction itself, most commonly based on the total amount of money bid. In Reiner Knizia's Amun-Re favor auction, the total amount of money sacrificed (bid) to the Pharoah determines how bountiful the land is. In Bruno Faidutti's Queen's Necklace jewel auction the total number of each gem bid determines the rarity of a gem and thus its value.

Bid Limitations

As with trading games, I think that limitations can really make an auction game, as players then have to jump their strategy through various hoops. Here's the main methods of limitation I've seen in auction games:

Victory Limitations: Most frequently, each player is only allowed to win one auction over a specific time period, effectively meaning they can't bid once they've won. This tends to be setup as follows: a number of items equal to the number of players are offered up for bid; they're bid on one at a time; but once a player has won one of the items, he can't bid again until a new set of items is offered up for bid. This creates some interesting dynamics where, by holding back, players can get less desirable items for very cheap prices. This method is found in Richard Ulrich & Wolfgang Kramer's The Princes of Florence and Reiner Knizia's Amun-Re land auction. A slightly different method is found in Reiner Knizia's Medici where each player is only allowed to win five auctions per turn, and where those auctions can be for goods with wildly different values, creating a real economy for high-value versus low-value cards.

Bidding Limitations: Players aren't allowed to bid unless they meet certain criteria--often having won a certain type of past auction. This is a central point of Tom Lehmann's Pizarro & Co. where you can bid for the services of various explorers in years I, II, and III, but you're only allowed to bid in years II or III if you already hold the same explorer's services from the previous year.

Monetary Limitations: Inevitably every game will limit in some way the monetary unit used for purchasing auction items. This is most frequently a limit on total resources, which may be supplemented from time-to-time in the game. Tom Lehmann's Pizarro & Co. further limits players by the exact bills they have in hand; they aren't allowed to make change and thus may have to spend more than they wish. Reiner Knizia's High Society uses this same concept and extends it: once you've put a bill down as part of a bid, you can't retrieve it, only supplement it.

Uniqueness Limitations: Some of the pseudo-auction systems don't actually require bidders to bid better than previous bidders; instead each bidder need only bid differently, then each of the Multiple Variable Winners gets an appropriate victory result. This is the case for the systems in Moon & Weissblum's New England and Kramer & Ulrich's El Grande, each of which requires each player to select a different "number" to offer as their auction bid.

Limitations must be approached with real care, because they have the potential to totally kick the bottom out of your auction system if they result in too few people bidding. Princes of Florence works great, because every round the number of people involved drops from N [N=3 to 5, the number of players] to 1--and then it jumps back up to N for the next round. Conversely, the auction system for Pizarro & Co. tends to degenerate. You have up to N [N=3 to 6] players involved in a year I auction, then 2 to 3 in a year II auction, then 1 to 2 in a year III auction. By the end of the game, auctions are fairly boring, and some players have no interaction at all with other players in the game.

Finally, there's one last auction idea that's related to limitations:

Bidding Repercussions: This isn't exactly a limitation, but closely related. What you bid can affect other portions of the game in negative ways. (Really, it's just a way to give your "money" value, like it would have in the real world.) For example, in Reiner Knizia's High Society, whoever has bid the most of his money by the end of the game automatically loses. In Reiner Knizia's Amun-Re the money used for bidding is used in two different types of auctions and also to purchase other items, and so your bidding has repercussions to a lot of other game systems.

Making a Good Auction Game

Having broken down auctions piece by piece, I now want to take a moment to offer my ideas on what makes a good auction game. First, the core auction must work:

Items Must Have Different Valuations for Different Players: This is my same core rule for trade games, and it applies here for different reasons. Technically, you could have items have the same value for all players, but then one of three things happens: either (1) the first player to jump-bid to the "right" price is the game winner -or- (2) the player who makes the best valuation of the items is the game winner -or (3) auction success is solely dependent upon other limitations, such as whether you have the right money units to bid. The aforementioned Pizarro & Co. is a game that seems to base its gameplay upon item valuation: the items up for bid have different valuations in very non-intuitive ways, and thus the gameplay doesn't turn out to be in the bidding, but rather in the figuring. Once you know the right formulas the whole game degenerates into boredom. Conversely, if items have different valuations for different players you open up whole new avenues of bluffing tactics. An item exceeds your personal valuation, but do you keep bidding because you think its value to your opponent might otherwise be too good?

If you get the bidding and valuations right, you can center a game pretty strongly on auction alone. A lot of the games I've mentioned do this, including Medici, High Society, and Pizarro & Co.. However, I think the best auction games do more.

Auction Games Can Have Other Systems: In the most complex and interesting auction games, the items you buy have other, continuing effects on your game. For example, in Corne van Moorsel's ZooSim you purchase "zoo tiles" which you lay out to form a zoo. Connecting your zoo tiles up in good manners to continue paths and keep together similar animals is a core component of the game. As already mentioned Reiner Knizia's Amun-Re has a lot of systems. You auction off Egyptian lands, then may hire farmers for those lands and build pyramids, then auction off Amun-Re's favor, which has results on a famer's output, then go back and buy more land.

A Few Examples

I've listed out multiple games throughout the article, but I'd like to close up by giving more complete examples of some of the more complex and, I think, better auction games.

Amun-Re: This Reiner Knizia auction game has two distinct auctions. In the first, players auction to purchase lands on the Nile. In the second, players bid for Amun-Re's favor.

The land auction is a fairly simple constrained English auction with open turn-based continuous bids. Each player, one at a time, puts their bid on a certain land, at a certain level. There are also both bidding and victory limitations. Once you've bid on a land, you must jump to another land if you're overvid; since you only have one bidding marker, by the end of the bidding each player will only be allowed a single victory. There's a first payer, single winner for each land.

The favor auction is an unconstrained English auction with simultaneous one-time blind bids. Each player simultaneously puts a bid into their hand, and they're all revealed at once. This is an all payer, multiple variable winners auction, where everyone puts out their money, and the first-place winner(s) get the best results, the second place the second-best results, and everyone else the third-best results. There is also a discrete result: the total valuation of the bid determines how productive all of the lands are that turn.

Queen's Necklace: This game also has two distinct auctions. In the first, players bid for gems, and in the second players try and sell the jewelery made of those gems.

The gem auction is a constrained Dutch auction with open turn-based continuous bids. Each player gets a set of 5 cards available to him for purchase with a set amount of money (10 ducats a turn). He takes his turn and buys any items he wishes, then the prices all decrease for the next player and new items appear to replace the items the previous player bought, each valued at their top price. There is a first payer, single winner for each auction, with no particular limitations.

The jewel auction is an unconstrained English auction with simultaneous one-time blind bids. Each player chooses a number of each gem that he will offer for sale. Victory is determined as all payer, multiple identical winners with discrete results. Everyone expends all the gems they bid, but only the players tied for the top values actually sell their gems, and that price can be decreased or increased by the total number of gems of that type sold. Note that this is effectively a seller's auction rather than the more typical buyer's auctions: the players are exchanging goods for money rather than money for goods.

El Grande: I offer this final game as an example because it's somewhat far removed from pure auction dynamics, but still clearly uses the same underlying systems.

In El Grande at the start of a turn you offer up a power card numbered between 1 and 13. This is a limited resource: you only have one of each card and nine rounds of play, and you don't get your cards back. The number determines how quickly you go in the round, and thus which of several limited actions you get to take. But, there's an additional balance: the higher the number you play, the fewer of your people you get to move to your "Court" (from which they can then be moved to the board). Thus, instead of paying with money like in most auctions, you're instead paying in lost opportunities, and instead of paying for goods, you're paying for new opportunities. Overall, other than the strangeness of the monetary units, it's an unconstrained English auction with open turn-based continuous bids. There are uniqueness, monetary, and victory limitations. Finally, there are all payer, multiple variable winners.


I've mentioned a number of different auction games in this article, all of which I've read the rules for except Merchants of Amsterdam, most of which I've played, and some of which I've reviewed. Here's RPGnet reviews that I've written so far: Amun-Re, El Grande, Fist of Dragonstones, High Society, New England, The Princes of Florence, Puerto Rico, and Queen's Necklace.

Although I've tried my best to include a good cross-section of auction games in this article, and I think I've hit almost all the major points, I've inevitable left some major games out. These include: many more auction games by Reiner Knizia, including Ra, Traumfabrik, and Taj Mahal; American "classics" such as Masterpiece; various railroad games involving bidding such as the "18xx" series; stranger auction games like Evo and Wildlife; and many more. Just pretend that your favorite auction game is listed herein.


Auctions are one of many methods that you can introduce into strategy games and roleplaying games alike. And, there's a lot more variability than just bidding around the table till everyone gives up. Through various methods of limiting information and limiting participation, in particular, you can introduce lots of strategy into auction games.

Finally, thanks to folks who offered particularly notable comments on this article, or on auctions in general, either at LiveJournal or spielfrieks, including David Goldfarb, W. Eric Martin, Doug Orleans, and Wei-Hwa Huang.

[ <— #143: If You Build It They Might Come: Critical Mass & Other User Buzzwords | #145: How to Cheat at Online Games —> ]

Miniseries ...

Recent Discussions on Trials, Triumphs & Trivialities: