Last post: Think twice about what you wrote. Today is the tomorrow of yesterday. Coming in today, one should expect the high to be broken and it did first! Now the low was expected not to be broken (this does not mean impossible) to break, if the trend was to continue). Since it was broken (and it was broken later in the day) it means that tomorrow (Monday) the expectation is that the high of today will be harder to break, and that the low should be easier to break. Moral of the story: I never wrote that both of them cannot be broken. You seem to think of higher high and high low as the upper trend. Up trend is uptrend until the low is broken. If low is broken, then trend is violated. Breaking the high is expected in up trend not needed, but is not what determines the trend.
I guess no need to publish your study because it is already comes with most software packages. Tree line break indicator is customizable to each individual. based on the close above/below previous preset close. or Donchian channel based on violation of previous preset high or low.
=============== Or to repeat helpful foundational pattern, which can be perhaps be built/ improved on; trend= friend. In other words, stocks can change personality, some times much more notable, with strong moves below/above 50 dma. An example of building on trend=friend, with 50 dma.
I was kind of skeptical but looks easy to test so I tried it with some S&P500 data I had lying around. It actually looks like it might work for the highs - I'll be damned. On days where you said the high was harder to break, here are the results: signal days other days time signal days high broken high broken -------------------- -------------- --------------- -------------- 1/3/00 - 4/5/07 847 40.1% 58.6% 1/3/06 - 4/5/07 135 44.4% 60.8% Signal days are where yesterday's low was lower than the low from 2 days ago. Lows were a little easier to break but not as dramatic.
I took another look and it is not as exciting as I thought. For 1/3/00 - 4/5/07 the average difference between yesterday's high and close was 11 pts on "signal days" and 5 pts on other days. The result was similar for 1/3/06 - 4/5/07 too. Of course it is harder for the market to break yesterday's high on a signal day, it has farther to go from yesterday's close. I also compared the open to high, open to close, and low to high ranges on days after the signal with other days, nothing interesting. Todays high-low range is slighty larger open to close returns are actually slightly higher on signal days, but not impressively. If that is all this signal is telling us about today's market I don't see how it is useful for trading the S&P500. Is this rule supposed to work on indexes? Is there supposed to be some effect at a different time frame? Is anybody reading this or am I just talking to myself?