Everything this video states about where orders rest, why obvious levels attract price, and what separates an accepted break from a rejected one, with the primary source for each.
The charts are invented and labelled Illustrative on screen. They teach the mechanism;
they are not evidence that any particular move happened, and nothing in the video claims a
win rate, a backtest result or a historical price.
On screen: about 10% of orders sat at rates ending in 00, against about 3% at each
other rate ending in 0.
The clustering was measured directly from a dealing bank’s complete order book rather than inferred from price. The sample is 9,655 stop-loss and take-profit orders with an aggregate face value over $55 billion, in dollar-yen, dollar-sterling and euro-dollar, placed between 1 August 1999 and 11 April 2000.
“Both stop-loss orders and take-profit orders tend to cluster at round numbers. Almost 10 percent of all such orders are placed at rates ending in 00 (such as ¥123.00/$ or $1.4300/£); on average, about 3 percent of orders are placed at each of the other rates ending in 0 … about 2 percent of orders are placed at each of the rates ending in 5.”
Carol L. Osler, Currency Orders and Exchange Rate Dynamics: An Explanation for the Predictive Success of Technical Analysis, Journal of Finance 58(5), 2003. Full text: https://faculty.georgetown.edu/evansm1/New%20Micro/osler1.pdf Federal Reserve Bank of New York Staff Report 125: https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr125.pdf
On screen: buy stops and the stops of shorts resting above resistance; sell stops and the stops of longs resting below support.
This is the asymmetry the same study isolates, and it is the reason the two pools in this video are drawn on opposite sides of their levels rather than symmetrically around them.
“Stop-loss buy orders cluster just above round numbers (meaning specifically numbers ending in 00 or 50), and stop-loss sell orders tend to cluster at rates just below round numbers. For example, 14.3 percent of executed stop-loss buy orders have requested execution rates ending in the range [01,10], while only 6.9 percent of those orders have requested execution rates ending in the range [90,99].”
Same source as above.
In the video: the cascade as price crosses the level and takes the resting orders, and the claim that a break can feed itself.
Stop-loss orders are positive-feedback: executing one pushes price further in the direction that triggers the next. The paper finds that price movements are unusually rapid where stop-loss orders are documented to cluster, and that the effect is larger for stop-loss orders than for take-profit orders, which are negative-feedback and therefore do not cascade in the same way.
Carol L. Osler, Stop-loss orders and price cascades in currency markets, Journal of International Money and Finance 24(2), 2005, pages 219–241. Federal Reserve Bank of New York Staff Report 150: https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr150.pdf
In the video: a large order failing to fill against a thin book, and filling at the level where the resting orders are.
A large order cannot be executed against a shallow book without moving the price against itself, which is why depth is the constraint rather than intent. The lower the depth, the faster the book must be replenished to absorb a given flow without prices moving, and large orders are therefore worked incrementally against available liquidity.
Jun Muranaga and Tokiko Shimizu, Market microstructure and market liquidity, Bank for International Settlements, Committee on the Global Financial System: https://www.bis.org/publ/cgfs11mura_a.pdf
Bouchaud, Farmer and others, The market impact of large trading orders: https://haas.berkeley.edu/wp-content/uploads/hiddenImpact13.pdf