Players can put up an offer to sell a fixed amount of a given commodity in exchange for a fixed amount of a different commodity.
Players may peruse the list of all open offers, and select one to "buy". Offers include the name of the player making the offer. "Compatible" offers are not matched up automatically. Offers are all-or-nothing so a large sell offer will have fewer potential purchasers; it will not be split into smaller transactions by the exchange.
Players have an upgradable limit to how many transactions they can participate in. A trade takes a certain amount of time to complete once agreed upon. If all your "merchants" are busy you cannot put up new offers or agree to existing offers until a previous offer or purchase completes.
The question: Without access to privileged data, what can an individual player do to monitor the "market"? An offer that disappears does not mean that a trade took place; it may simply mean that the offer was withdrawn.
Obviously a player can try to judge what the market will bear by raising and lowering his own prices. But, since the transaction cost is high (limited number of trades), there is significant benefit to being able to determine a good selling price without repeated experiment.
The simplest way of analyzing the data is to take it at face value--- if you see a player offering 40X for 30Y, then a counter-offer for 30Y for 40X will have at least one customer who thinks that the offered price is fair.
Another way thinking about the situation is to collect "negative" information. An offer that remains listed for a long period of time is obviously not offering a price that purchasers find reasonable. Unfortunately this may be difficult to analyze because "large" offers may offer an attractive rate of exchange but at a quantity which is only available to a small portion of the players. But we can put transactions into different buckets and compute price bounds for each size.
A final idea is that offers which are re-listed by the same player may indicate successful trades. A player who successfully sold 100X for 125Y is unlikely to make his next offer 100X for only 75Y. Since the offers are not anonymous we could compute a time series for a particular player. (A quick re-listing at a lower price, of course, may indicate that the player has lowered his estimate of the commodity's value in an attempt to generate a sale.)