Inferring Trader Identity from Limit-Order-Book Data
Summary
The document asks whether liquidity suppliers can determine who initiated trades from Nasdaq ITCH data. The accepted answer says ITCH exposes match numbers and displayed-order reference numbers, but no buyer-side identity. These fields support trade correction and order tracking rather than identifying the aggressor.
Other answers describe indirect inference from event sequencing, timestamps, and order sizes. A researcher might group executions that arrive together, study repeated size patterns, or track a participant’s apparent response to price changes. One answer claims this approach identified recurring participants in CME futures data, but the evidence is anecdotal and depends on precise knowledge of exchange matching and message dissemination. Similar timing or sizes do not prove that orders came from the same trader. The document also notes that some Nordic equity trade messages may include buyer and seller participant IDs, so identification depends on the venue and feed specification.
Key ideas
- Nasdaq ITCH does not include a direct identity field for the buy-side aggressor.
- Match numbers and displayed-order references serve trade and book management purposes, not trader identification.
- Message sequencing, timestamps, and repeated order sizes can support indirect participant inference.
- Such inference relies on exchange-specific behavior and does not establish identity with certainty.
- Some Nordic equity feeds may provide buyer and seller participant IDs.
Tags
Full text
# Trader's identity in a limit book # Trader's identity in a limit book In a limit book like NASDAQ ITCH, can liquidity suppliers know the demand-side identity of a trader prior or after a trade? Knowing this will help me with my theoretical model that I am trying to develop since I am wondering whether a liquidity supplier that observes say a trade of 90 shares that takes away liquidity up to three ticks (30 shares per tick - block shape book) can distinguished whether the trade came from one individual or three individuals who submitted trades at each tick. ## Answer by chrisaycock (score 3, accepted) https://quant.stackexchange.com/a/8978 ITCH does not disseminate any identifiers for the buy-side. They have a match number (used to correct or break trades) and a reference number (for displayed liquidity) and that's it. No other identifying features are present. ## Answer by Serg (score 3) https://quant.stackexchange.com/a/9029 I was able to identify significant participants by order size on CME exchange. I think ITCH is even more informative that CME's data format. The trick is to learn very closely the incremental data and the order in which this data arrives. We can assume that exchange's Matching Engine and its market data distribution algorithm are programmed machines, therefore always process similar cases in exactly the same way. For example, a large aggressor order is being executed. First, you receive all the matched executions. Then you receive order book updates, caused by this executions (In fact, knowing the matching algorithm, you can predict the book updates!). Even if there is another aggressor order waiting, its executions will not be transmitted until the previous order book updates are completed. Of cause, you need to validate this with timestamps as said above. But there are much more special cases, which I can't describe here. For example, the order may be only partially executed until reaches its limit price. It can still be detected. Such investigation can be applied on both aggressor orders and passive limit orders activity. Below is such an example on one day of ES future. For example, there are 11285 events of BUY 46, and just ~200 events of BUY 47. Which makes it reasonable to assume that most of orders of size 46 belong to the same trader. Some traders attempt to "hide" their activity by random distribution (between 100 and 120 in this example). Also as expected there are peaks on round sizes 100, 200, ... After detecting distinguishing peaks of order size, you can separate certain trader to watch its behavior. The second image illustrates the activity of trader who uses order size 46. We can see its behavior before news release. It's also possible to detect his location using response time to price changes. This trader's machine is located in Europe because its response time is about 100 ms. ## Answer by LazyCat (score 1) https://quant.stackexchange.com/a/8984 Regarding the second part of your question, - if you have relatively precise timestamps, you can use those to distinguish the cases you're interested. E.g. if one party took all three levels, the timestamps will be very close, or identical. ## Answer by da_m_n (score 0) https://quant.stackexchange.com/a/15150 AFAIK Nasdaq OMX is one of the few (if not the only) exchange which provides buyer / seller ID on equity trades: > Trade information (latest paid price, trade volume, buyer/seller ID where available) http://www.nasdaqomx.com/digitalAssets/66/66627_nasdaq_omx_nordic_market_data_offering.pdf More specifically: > TRADE MESSAGE (NON-CROSS) Name Offset Length Value Notes ... Participant ID, buyer 45 4 Alphabetic Participant ID, seller 49 4 Alphabetic http://www.nasdaqomx.com/digitalAssets/94/94202_nordic-equity-totalview-itch-1.90.4.pdf
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.