Read the intermarket weight matrix first, then the predicted moving-average filter

A two-layer chart read

A 2018 product case study described searching daily observations from thousands of global markets to isolate thirty related markets with the greatest influence on a chosen target, then encoding that influence as a numeric intermarket weight matrix.

The same write-up also showed predicted moving averages, predicted differences, a predicted next-session high and low range, and a typical-price forecast. Editorial: read those pieces as two layers. First collapse the global book into the fixed matrix. Then treat the predicted moving-average stack as a short-horizon trend filter on typical price rather than as an entry ticket.

The intermarket weight matrix

The write-up contrasted that multi-market weight construction with computing linear correlations between only two markets at a time.

An intermarket weight matrix is a fixed set of related markets, here thirty, whose influence on one target is stored as numeric weights instead of a single pairwise correlation.

Predicted moving averages and typical price

The visible indicator set included predicted short-term, medium-term, and long-term moving-average crossovers; predicted short-term, medium-term, and long-term differences; and a predicted next-session high and low range.

A predicted moving average is a moving-average line whose level is generated from a short-horizon forecast rather than from completed prices alone.

One undisclosed index compared the current three-day moving average with a predicted three-day moving average to forecast whether typical price would be higher or lower two days later. A typical-price forecast is a two-day-ahead statement of whether the average of high, low, and close is expected to be higher or lower.

A trend filter, not a ticket

An accompanying chart used green shading for bullish trend segments and red shading for bearish segments when illustrating forecasted trend direction from predictive moving averages.

Editorial: treat that shading as a trend filter, a directional overlay that marks bullish versus bearish segments without issuing an entry or exit instruction.

The software was described as not generating buy or sell instructions. Users were expected to combine the near-term directional output with their own reading of the target market.

Coverage and a detached market scan

Listed coverage included close to 2,000 symbols spanning futures groups, 13 major and 21 cross currency pairs, multiple ETF categories, and US and Canadian stocks, plus near-term movement forecasts for bitcoin, Ethereum, and Litecoin.

Forecast calculations were moved to a remote server so chart data appeared when a chart was opened. A market-scan function was detached from portfolios and charts so multiple scans could be stored independently. A market scan is a saved, chart-independent screen that ranks a chosen universe by the same predicted indicators.

How the write-up described the patents

The case study cited a 2013 patent on locating related intermarkets and a second patent on converting lagging indicators into leading indicators intended to forecast trend and trend change a few days ahead.

The second patent is the leading transformation in this article: a claimed recast of a lagging average into a few-days-ahead reading of trend and trend change.