This is a plain old missunderstanding of the concept of movement. As I said, only unfilled gaps classify as overnight movements, any gap filled during the day means that the actual move also happened intraday. Example: SPX opens with a 20 points upgap, but then drops back and fills it 75%, before continuing the rally. So how much movement was achieved overnight really? A whole 5 points!!! 3/4 of the movement was also and eventually done intraday! (and I don't even bring up the subject of gaps filled a few days/weeks later) Last year the SPX dropped 560 or so points. I bet there is no more than 25 points gap left behind which is below the 5%, what I stated in my previous post....
You are on the right track but what you need to be analyzing is not what the S&P has done in the last 100 years but in the last 6-18 months. Lets say in the last 6 months the market has been in an uptrend. The probability that the market will open higher is much greater than opening lower based on previous statistics of your own analysis when it comes to the uptrend. The problem with a 100 years is that even though the market has gone up for the last 100 years if you would have bought and sold the open in the last year you would have to be out of money. If you can adapt to the market than you can figure out which probabilities are in your favor and hopefully make profits if you can manage not to pysch yourself out... There are other things you can do as well instead of just focusing on the open. For example if you notice through analysis that most of the strongest movements are in the middle of the day than you should adjust your strategy to the middle of the day instead of just faithfully staying with the open.
Perhaps I shouldn't have used the term movement, as my analysis did not identify every intraday movement of the market, but merely the cash open, cash close, and the corresponding globex values. So in your example SPX opens 20 points up, falls off and closes the gap, then rallies 20 points back up to close where it opened, my analysis would say that was a 20 point overnight movement, since no value actually changed during the day from the open to the close. From what I can tell, you are saying this 20 point gain can be argued to have been either achieved overnight, or during the day, since the same 20 point prices were traveled over during both sessions. But the 20 pt gain has to be attributed to one of them. I am simply using OPEN/CLOSE, where you seem to be interested in HIGH/LOW vs unfilled gaps or some kind of range calculation. Please correct me if I'm wrong. In my example above, where would you attribute the 20 point gain to, as it happened in both markets?
The overall market trend has little to no relationship with the intraday trend. In a bull or bear market, overnight action (gaps) make up most of the move. In terms of intraday, there is no trend. Added up over many hundreds of days, the sum difference between open and close is nil. Holding overnight is where the trend-following money is made. I did a study of Google when it was in its bull run (around two years' worth of data). Buying the close and selling the next day's open made a trader an average around $1 a day. Buying a day's open and selling that same day's close netted nothing. The same holds for the S&P. Of course there are fluctuations throughout the day, and it is by taking advantage of those are where the intraday trader makes a living. I use floor trader pivots and trade off the open every day (I consider myself a tape reader in the classic, pre-T&S version). But my point was that an intraday trader cannot simply go long just because the end of day trend is up. Intraday, overall, there IS no bias. The sum total of the intraday fluctuations is zero; intraday trading is a zero sum game minus vig, which at least partially explains why so many intraday traders fail and should probably stick to longer-term, end-of-day trading. For the most part, price hangs around pretty close to the open. If price gets too far from the open, expect a reversal, being aware that on some days, like yesterday, there WILL be a trend. One can easily cherry pick periods in which these principles do not work as well. But OVERALL, in the LONG HAUL, they are true.
I would attribute it to the day's movement, because it did go back closing the gap. So at that point the trader who went long the previous day was back at breakeven. If you look at a daily chart that shows only the day's movement you will see very few unclosed gaps in a year...