How Public Order Books Inform Trading and Manipulation Strategies

# Orderbook Arbitrage


# Orderbook Arbitrage












The order-books of trading exchanges are often hidden as so-called "Dark Pools". The measure was taken to avoid apparent market manipulation strategies executed by traders back then.

Which such arbitrage/manipulation strategies are possible if the order book is public?

Explain the strategies in detail.

Hint #1: There are three example strategies in the comments below.

Hint #2: Google.

## Answer by boot4life (score 14, accepted)

https://quant.stackexchange.com/a/18876

A public order book gives traders information not only on the current price of a security, but also the volume and structure of the entire supply and demand schedule.

Such information can be used for arbitrage and market manipulation strategies in various ways:













## Answer by lehalle (score 10)

https://quant.stackexchange.com/a/18761

I am not sure Dark Pools (DP) have been created to avoid "market manipulation". They have been created by firms because they found an advantage to create them (see Market Microstructure in Practice, L and Laruelle Eds.). The main reasons have been:

- spare market fees, for DP created by brokers (like UBS MTF);

- spare market impact, for block pools (like ITG/POSIT);

- spare half of the bid ask spread, for DP created by market makers (like Knight Link).

Optimal trading in DP is possible, see for instance Optimal split of orders across liquidity pools: a stochastic algorithm approach, by Laruelle, Pagès and L (published in SIAM Journal on Financial Mathematics, Vol. 2 (2011), pp. 1042-1076). You can have a look at Optimal Allocation Strategies for the Dark Pool Problem, by Agarwal, Bartlett, and Dama. And to Censored exploration and the dark pool problem, by Ganchev, Nevmyvaka, Kearns, and Vaughan too.

In terms of optimal trading in orderbooks, you have few nice papers:

- one for very short term trading: Optimal posting price of limit orders: learning by trading, by Laruelle, Pagès and L (again), published in Mathematics and Financial Economics, Vol. 7, No. 3. (11 June 2013), pp. 359-403.

- one to make the market: Dealing with the inventory risk: a solution to the market making problem, by Guéant, Fernandez-Tapia and L, in Mathematics and Financial Economics, Vol. 4, No. 7. (3 September 2013), pp. 477-507.



You will find here all that you need to build you own orderbook strategy.

I strongly suggest you have a look at other posts on Quant.SE, since you will need to understand:

- how to backtest your strategies: How to design a custom equity backtester?

- how to include latency into it: What latency should I use for backtesting a high-frequency strategy?

EDIT: these strategies are mostly robust to gaming. As soon as you value your position in terms of price and risk (i.e. using a value function) and not relatively to the mid, it is far more difficult to game you. Of course you can suffer from adverse selection, but only conditionaly to the fact you agreed on the price with respect to an historical measure. You can add some usual anti gaming features, like not using the same quantity for each of your orders, but again if your sizes are optimized according to a sophisticated enough measure, you will never place two orders with the same size.