Maintaining a Target Share of Limit Order Book Queues
Summary
The document discusses how a market maker might maintain a target fraction of displayed depth at each price level. The proposed rule scales the trader’s order with other participants’ orders: add a proportional amount when new depth arrives, and reduce the trader’s lowest-priority quantity when others cancel. When market orders consume liquidity, adjust orders according to whether the market maker’s own quotes execute.
The example illustrates the mechanics of tracking a roughly uniform share of each queue, but it is not a complete implementation guide. The author stresses that uniform queue representation is a simplifying assumption used in a model, rather than necessarily the best practical objective. Market makers generally value priority near the front of the queue. A follow-up raises a risk trade-off: concentrating orders at the front may make several execute together during a large trade, while spreading them out may reduce that exposure but affect execution likelihood. No empirical comparison resolves which placement is preferable.
Key ideas
- A target queue share can be approximated by scaling additions and cancellations in proportion to observed depth changes.
- When market orders arrive, quote adjustments depend on whether the market maker receives executions.
- Maintaining a uniform share across price levels is presented as a modeling simplification.
- Queue priority favors placing orders near the front, while clustering can increase exposure to large trades.
- The discussion offers no evidence establishing one queue distribution as universally safer or more profitable.
Tags
Full text
# Achieving an even distribution of orders in the queue # Achieving an even distribution of orders in the queue Baron Law, Frederi Viens: Market Making under a Weakly Consistent Limit Order Book Model contains the following paragraph > "The market maker may achieve her target execution profile by continuously adjusting her limit orders in the LOB to be roughly the proportion ρ of the queue length at each price level, but the detailed mechanism is outside the scope of this paper." Does anyone have any references for how this is achieved in practice given the presence of both trades and cancels removing depth at the best prices? ## Answer by Baron (score 4, accepted) https://quant.stackexchange.com/a/46009 Suppose your target participation rate is 1/11 ~ 9%. At each price level, whenever someone puts a limit order of size 10, you put a limit order of size 1 right after him. Whenever someone cancel an order of size 10, you cancel 1 share from your limit order with the lowest priority. When a market order arrives, you adjust your limit orders depending on whether your limit orders are executed or not. However, I would like to stress that the uniform distribution assumption is more like a simplifying assumption, rather than the goal, in deriving the optimal risk limit in my paper. As a market-maker, you always want your limit orders to be at the top of the queue. ## Answer by EdisonKIng (score 0) https://quant.stackexchange.com/a/77250 I also wonder if targeting uniform distribution is better than clustering at the top of the queue in some cases. For example, if bid orders are distributed uniformly, when a bid order is filled, the next bid order will be filled at a short time with relatively low probability. In the meanwhile, if the ask order of the market maker with the same size is filled, then the market maker’s position is stilled zero. Instead, if all bid orders are clustering at the top, they may be filled together by a large market sell order, resulting in a large risk exposure with long position. In the sense of risk, is uniform distribution better, though it is difficult to achieve.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.