Signing Off-Exchange And Odd-Lot Prints

A large and growing share of US equity volume executes away from the lit exchanges — in dark pools, wholesaler internalization, and other off-exchange venues — and gets reported to the consolidated tape with a delay and with venue codes that differ from a normal exchange print. Odd lots (trades smaller than the standard 100-share round lot) are reported too, but historically weren't even included in the official best bid/offer calculation or in some volume statistics, despite representing a real and, for high-priced stocks, increasingly common share of trading.

Both quirks matter for anyone reconstructing signed order flow from the tape. Off-exchange prints often reflect retail flow being internalized by a wholesaler at the midpoint, which behaves differently from lit-market flow — it's typically less informed and executes at a fixed relationship to the prevailing quote rather than by crossing the spread, which can break trade-classification rules like the tick test or EMO that were built assuming an exchange limit order book. Odd lots skew toward algorithmic slicing of larger parent orders and toward high-priced stocks where round lots represent large dollar amounts, so ignoring them (as older data feeds did) understates true volume and biases order-flow statistics for exactly the securities where odd-lot trading is heaviest.

A careful order-flow study needs to flag venue and lot-size fields explicitly, decide whether to include or separately analyze off-exchange and odd-lot prints, and be aware that any historical dataset predating the SEC's later odd-lot reporting reforms will simply be missing this volume outright.