Multi-Timeframe Darvas Box Analyzer
Nicolas Darvas built his box in How I Made $2,000,000 in the Stock Market (1960) from a high that 3 later bars couldn't clear and a low that 3 later bars couldn't break after it. That rule works on any bar size, so this analyzer builds one box on your daily bars and a second box on the same data rolled into weekly bars, then checks every time the daily box broke a wall while sitting nested inside a live weekly box: did the close also clear the weekly wall, or did it snap back inside? Paste bars with a leading date and the weekly box uses calendar weeks; paste bars without one and it groups every 5 bars into a week instead. The read tells you which pattern this instrument has actually favored, then applies it to the box sitting in front of you now.
One bar per line, oldest first. A leading YYYY-MM-DD date groups the bars into calendar weeks for the second box; skip the date and every 5 bars becomes a week instead. Paste 60 bars minimum so both boxes have room to set; 120 or more gives the nested-break count something worth trusting.
No boxes yet
Paste daily bars, then hit Build. Or load an example to see a range regime where nested boxes fade back in, a trend regime where they run through, and a short paste with no dates.
How each box builds. Same Darvas rule on both series: a top is a high that 3 following bars fail to clear, a bottom is the lowest low since that top that 3 following bars fail to break, and the box ends on the first close outside either wall. The weekly series is built by rolling the daily bars into calendar weeks (Monday start) when you paste dates, or into blocks of 5 bars when you don't. The backtest. For every daily box that broke a wall while it sat fully inside a set weekly box (its top at or under the weekly top, its bottom at or over the weekly bottom), the analyzer checks whether that same closing bar also cleared the weekly wall on the same side. A close that stayed on the weekly side of that wall counts as faded; a close that cleared it too counts as confirmed. Nested breaks need at least 3 resolved cases before the rate counts. The live read. If the current daily box sits inside the current weekly box, the fade plan enters at the last close, stops 20% of the daily box width past the near wall, and targets the midpoint then the far wall, same as fading any Darvas box. Keyboard: Ctrl+Enter builds, 1 to 3 load the examples, C copies the read.
Educational tool only, not financial advice. The nested-break rate comes from your pasted history. A short paste or a thin instrument won't leave enough resolved breaks to trust the number, and every rate can flip once the regime changes.
Built by Jason Parker, founder of Trading Ranges.
Built by Jason Parker, founder of Trading Ranges.