A Rule 605 execution quality report is a monthly public file in which a trading venue or a large broker discloses how its market orders and marketable limit orders in US stocks were filled: how often at the quote or better, how much price improvement they received, how wide the effective spread was, and how quickly fills came. Under the SEC's amended Rule 605, whose compliance date moved to August 1, 2026, brokers with 100,000 or more customer accounts now have to publish these reports, and the first ones, covering August 2026, were due by the end of September 2026. As of October 2026 they are online.
Rule 605 of Regulation NMS requires market centers, such as exchanges, alternative trading systems and wholesalers that fill retail orders, to publish monthly statistics about the orders they handle. In its March 6, 2024 adopting release (Release No. 34-99679), the SEC extended that duty to brokers and dealers that introduce or carry 100,000 or more customer accounts and added a second, shorter document, the summary report. The SEC later moved the compliance date from December 14, 2025 to August 1, 2026 in a September 2025 order.
Each reporting entity now files two products. The detailed report breaks results down by symbol, order type and order size and covers more kinds of orders, while the summary report under Rule 605(a)(2) covers only market and marketable limit orders and comes as a CSV file plus a PDF built from the same data. Reports are due within one month after the end of the month they describe, so the SEC said August 2026 reports could be posted through the end of September 2026, and each must stay online for three years.
A monthly report describes a firm's whole book of orders, not your fill. If your question is why one order filled where it did, hi2morrow's guide to why a stock order gets a bad fill covers that; this guide is about reading the report itself.
On the free website where the firm, or a vendor or exchange acting for it, posts its files. Under the Rule 605 NMS Plan every reporting entity designates a self-regulatory organization, and that organization keeps a page listing the firm's reporter code and a link to its reports, as the SEC staff's Rule 605 FAQ explains. FINRA's page of Rule 605 reporters lists the firms that chose FINRA, and as of October 2026 it shows 124 broker-dealers alongside market makers and trading systems. Firms that picked an exchange appear on that exchange's page instead.
File names follow the SEC's technical specification for the summary report: the reporter code followed by the year and month, so NYSE's report for August 2026 is XNYS202608. A broker that is not itself a market center and has fewer than 100,000 customer accounts is not required to publish one, so a missing report from a small broker is not by itself a warning sign.
Every row is one combination of three things: whether the stock was in the S&P 500 on the first day of the month, whether the orders were market (M) or marketable limit (ML), and a dollar order-size band. The bands run from under $250 through $250, $1,000, $5,000, $10,000, $20,000 and $50,000 up to $200,000, then $200,000 or more, and a ninth row combines every order below $200,000. In the CSV each band is coded by its lower edge, such as 0, 250, 1000 or 50000, but the combined row carries the code 199999, which is easy to mistake for a $50,000-to-$200,000 band when it is actually the total for everything under $200,000.
Order size is measured in dollars at the quote that stood when the order arrived. According to the SEC staff FAQ, a buy is valued at shares times the national best offer and a sell at shares times the national best bid, and marketable limit orders are valued the same way as market orders. If a firm had orders in a combination but no executions, the statistics are left empty, and a combination with no orders gets no row.
NYSE's August 2026 summary file shows the layout: two sections, two order types and nine size rows give 36 data rows. Its first row covers S&P 500 stocks, market orders under $250, with an average order of 1.55 shares, an average notional size of $123.62 and an average midpoint of $79.53. The remaining columns are the execution-quality statistics explained below; the row illustrates the format and is not an assessment of NYSE or anyone else.
Each of these compares the execution price with the national best bid and offer (NBBO) at the moment the order arrived. A buy executed at the best offer was executed at the quote, a buy executed below it received price improvement, and a buy executed above it was executed outside the quote; sells mirror this against the best bid. The column "percentage of shares executed at the quote or better" counts shares at the quote plus shares with price improvement, so it is a share count, not an order count.
Share-weighted average percentage price improvement nets the two sides. Rule 605(a)(2)(vi) takes the cumulative amount that prices were improved, subtracts the cumulative amount that prices were outside the quote, and divides by the average midpoint times the shares executed, which means the figure can be negative.
The spread columns use the midpoint of the NBBO. The effective spread is double the gap between the execution price and that midpoint, the quoted spread is the best offer minus the best bid at arrival, and the percentage versions divide by the average midpoint. E/Q, or average effective spread divided by average quoted spread, is 100% for a fill exactly at the quote, lower when fills tend to land inside the spread and higher when they tend to land outside it.
Take ABC, a hypothetical stock outside the S&P 500, at 10:05 a.m. ET with an NBBO of $49.98 by $50.00, so the midpoint is $49.99 and the quoted spread is $0.02. A broker receives three market orders to buy 90 shares each, each valued at 90 times $50.00, or $4,500, which puts them in the $1,000-to-$5,000 band. The first fills at $49.995, an improvement of $0.005 a share and an effective spread of $0.01. The second fills at $50.00, at the quote, with an effective spread of $0.02. The third fills at $50.01 after the quote moved, outside the quote by $0.01 a share, with an effective spread of $0.04.
If those were the only orders in that row for the month, 180 of 270 shares were executed at the quote or better, which is 66.67% and appears in the CSV as 0.666667, and 90 shares received price improvement, which is 33.33% or 0.333333. Net price improvement is $0.45 of improvement minus $0.90 outside the quote, or minus $0.45, divided by $49.99 times 270 shares, which gives about minus 0.0033% and a CSV value of -0.000033. The average effective spread is $0.02333, or 0.0467% of the midpoint, and the quoted spread is 0.0400%, so E/Q is 116.67%. If the fills took 40, 60 and 500 milliseconds, the share-weighted average speed is 200 milliseconds.
The same row therefore shows two-thirds of shares executed at the quote or better and, on average, a cost higher than crossing the quoted spread. One poor fill in a fast market is enough to move an aggregate this way, which is why a row has to be read as a whole.
Execution speed is the share-weighted average time from order receipt to execution, in milliseconds. The unit matters: NYSE's first row shows 0.420, meaning well under a millisecond, while other rows in the same file run into the thousands, so always check the unit before comparing two numbers.
Realized spread compares the execution price with the NBBO midpoint a set time after the trade. The summary report gives the average percentage realized spread at 15 seconds and at one minute after execution, following the definition in Rule 600(b). Like the other spread columns it can be positive or negative; the detailed report adds 50-millisecond, one-second and five-minute horizons.
The preamble to the rule itself says the statistics alone do not create a reliable basis to judge whether a broker failed to obtain the most favorable terms available for customer orders. The numbers pool many customers, symbols and market conditions. For your own orders, Rule 606(b)(1) lets you ask your broker where they were routed over the previous six months, and hi2morrow's guide to why a market order fills at a different price explains how to test a single fill.
Some orders are left out. Under the SEC's April 2026 exemptive order, marketable orders received during an announced halt, or less than five minutes before one and still open, are excluded, as are orders affected by an NBBO crossed for 30 seconds or more and stocks averaging no more than five reported trades a day for the previous six months. Because the summary covers only market and marketable limit orders, it says nothing about resting limit orders or other order types, which appear only in the detailed report.
Price improvement relative to the best displayed price, which includes odd-lot quotes, is not in the summary report at all. It belongs to the detailed report, and per that September 2025 order it is required beginning November 2026; the SEC FAQ says November reports are due by the end of December 2026. In NYSE's August 2026 detailed file, the best-displayed-price columns are empty in all 403,113 rows, which is consistent with that timeline.
The practical traps are about format. The CSV stores fractions, so NYSE's 0.937238 for shares at the quote or better appears as 93.72% in the PDF and its E/Q of 0.405582 appears as 40.56%, even though the SEC specification describes E/Q as a percentage. Check whether a file uses fractions or percentages before comparing anything. Do not compare different size bands, order types or the S&P 500 and other-stocks sections with each other, because they describe different markets, and do not read a trend into one month.
Alexander Styopin's professional view: the useful way to use a summary report is to find the one row that matches how you trade, meaning the order type you send and the dollar size you usually send, and read "at the quote or better" together with E/Q rather than either alone. A high share of fills at the quote can sit next to an E/Q above 100% if the rest filled badly, as the ABC row shows. Then watch that row over several months, because a single month mostly records what the market did.
Educational material only. Not investment or legal advice. Order handling, margin, and settlement rules can vary by broker, account type, and jurisdiction.
Author: Alexander Styopin, hi2morrow analyst and economist with 25 years of experience in the US stock market