Where the ideas in this video come from, and what to read if you want to go further than nine minutes allows.
VWAP is unusual among the things covered on this channel in that it did not start life as a chart indicator at all. It was a way of scoring a broker, and it arrived on retail charts roughly twenty years after institutions began being measured against it. Most of what follows is about that history, because it explains why the line behaves the way it does.
The first execution measured against VWAP is attributed to James Elkins, then head trader at the New York agency brokerage Abel Noser, who used it in 1984 for the Ford Motor Company pension fund. The idea was not to predict anything. It was to answer a question a pension fund could not otherwise answer: given that our broker spent all day buying, did they buy well?
The academic formalisation came four years later. Berkowitz, Logue and Noser used the volume weighted average price over the trading day as the yardstick for total transaction cost on the New York Stock Exchange, across roughly fourteen thousand institutional trades. The Noser on that paper is the Noser of Abel Noser, which is the connection between the two entries above.
That paper is the source for the video’s central point about why large orders care: the line is a scorecard on a job, and the people being scored by it are the people whose orders are big enough to move price.
VWAP is the cumulative value traded divided by the cumulative volume traded over the chosen window, with each bar contributing its typical price weighted by the size that went through it. Because it is cumulative from the session open rather than rolling over a fixed number of bars, it resets every session, which is why it describes today and nothing longer.
The standard deviation band idea is not native to VWAP. It is John Bollinger’s, developed in the early 1980s, and its original form puts the bands around a simple moving average. Applying the same construction with VWAP as the middle line instead is a later adaptation, and it is what the video’s orange bands are.
The distances quoted in the video — about two thirds of the action inside the first pair, about ninety five per cent inside the second — are the empirical rule for a normal distribution (68/95/99.7), which is where every platform’s default band settings come from. Worth knowing that real price returns have fatter tails than a normal distribution does, so the third band gets touched rather more often than 99.7 per cent would suggest.
The observation that the bands pinching in tends to precede a fast move is Bollinger’s own, and he named it the squeeze. His formulation is that low volatility follows high and high follows low, so a contraction indicates a pending expansion without indicating its direction.
The video’s test for a real break of the middle line — did price stay there, or did it dip under and come straight back — is Market Profile’s idea of acceptance, and that is J. Peter Steidlmayer’s, developed in the early 1980s at the Chicago Board of Trade. The underlying claim is that the time a market spends at a price is what tells you whether the price was accepted or rejected, and a price the market keeps trading at is a price it has agreed to.
The video’s warning about the line going quiet is a real and well documented daily pattern rather than an impression. Intraday volume is U shaped: heavy at the open, falling to a minimum around the middle of the session, and rising again into the close.
The scorecard use described in the video is still current, and it is embedded in regulation. Article 27 of MiFID II requires firms to take all sufficient steps to obtain the best possible result for clients, and transaction cost analysis against benchmarks — VWAP among them, alongside implementation shortfall and arrival price — is how firms evidence it.
The video gives rough figures for how often a fade from the band returns to the middle on a rotating day against a trending one. These are the figures the trading education literature generally quotes, and they should be read as the shape of the difference rather than as measured constants: the gap between the two regimes is the durable finding, and the exact proportions vary by market, by session and by how the two day types are defined in the first place. Published backtests of band fades with a regime filter tend to land in the region of 55 to 65 per cent, and materially lower with no filter at all, which is the same point the video is making about reading the day before taking the setup.
Anyone wanting to hold themselves to a number here should measure it on their own market and their own definition of a rotating day rather than adopting one from a video, including this one.