1: \begin{abstract}

2:   Ensuring sufficient liquidity is one of the key challenges for

3:   designers of prediction markets. Various market making algorithms

4:   have been proposed in the literature and deployed in practice, but

5:   there has been little effort to evaluate their benefits and

6:   disadvantages in a systematic manner. We introduce a novel

7:   experimental design for comparing market structures in live trading

8:   that ensures fair comparison between two different microstructures

9:   with the \emph{same} trading population. Participants trade on

10:   outcomes related to a two-dimensional random walk that they observe

11:   on their computer screens. They can simultaneously trade in two

12:   markets, corresponding to the independent horizontal and vertical

13:   random walks. We use this experimental design to compare the popular

14:   inventory-based logarithmic market scoring rule (LMSR) market maker

15:   and a new information based Bayesian market maker (BMM).  Our

16:   experiments reveal that BMM can offer significant benefits in terms

17:   of price stability and expected loss when controlling for liquidity;

18:   the caveat is that, unlike LMSR, BMM does not guarantee bounded

19:   loss. Our investigation also elucidates some general properties of

20:   market makers in prediction markets. In particular, there is an

21:   inherent tradeoff between adaptability to market shocks and

22:   convergence during market equilibrium.

23: 

24: 

25: \end{abstract}

26: