In this case, bad measurements are OK. They average out over time. If the bias is one way, then one type of agent (seller or buyer) gets a bit more pay than the

In this case, bad measurements are OK. They average out over time. If the bias is one way, then one type of agent (seller or buyer) gets a bit more pay than the other, but this is not a disaster and is compensated by agents competing to work on the higher-paying side of the transactions.