%% Maybe in some markets; sounds like a big error to assume that is true in SPY,QQQ. NO way is SPY,QQQ close price,or open price to close price , negative /down past 20 years
Have you looked at the data sets and graphs that have been posted? It is 100% true, it's right there in the data. At the time of that article, in SPY the O2C session net gains were -4%. In QQQ it was much worse with the O2C session down -65%.
Are weekends included in the overnight study? Do weekends have more effect than any of the five periods. Buy Mon. Close - sell Tues open. Buy Tues close - sell Wed open. Buy Wed close - sell Thurs open. Buy Thurs close - sell Fri open. Buy Friday close - sell Mon open. .....weekends
Sounds about right, most up moves are gaps and market generally sells back a bit before close. Gaps are a way to manipulate the price on low volume, why i stopped trading stocks when forced to hold over night.
Just a note on this data set, 03/05/2015 and 03/30/2015 have an issue where the low of the day is higher than the open. It isn't going to cause this test any issues since I am not using highs/lows, but could for someone else, so just an FYI.
https://www.nytimes.com/2018/02/02/your-money/stock-market-after-hours-trading.html "His data shows that during the bear market year of 2008, the overall market, as represented by the SPY E.T.F., declined 36.8 percent. But most of the damage occurred during the day, with losses of 26.7 percent, compared with only 13.8 percent overnight." Since the graph in the article depicts cumulative returns, intuitively I would think the 2008 drop in the "regular hours" graph should be bigger than the 2008 drop in the "outside regular hours" graph. However that is clearly not the case. In a cumulative returns graph I would think a given percentage point change would look the same irrespective of the starting point? I will try to replicate the data when I have more time to find out why my intuition fails.
The main problem with trying to do this now is risk of ruin. You are assuming positive surprises that happened in the past will continue into the future. This is why most simple automated systems fail. For example, a negative black swam occurs and you have no stop loss. Now you might say this is impossible, but look at the $ millions that were made by traders selling volatility to lose 80% of their net wealth and client's money due to a spike in vol. I do trade at night since I wake up late. However, I currently use a stop loss and profit target. Most of my trades are designed to finish before a news event takes place.
And not true. If it were true, the charts would be full of UNFILLED upgaps. There are few such upgaps in the last 2-3 decades, but they are rather the exception than the rule. If the market upgaps, but during the day it fills the gap (thus comes back to the previous close) and rallies again, then the real gain happens DURING regular hours. Simple as it is. Just think about this: People would notice such regularly occurring things and would start to buy at the end of day and sell just after the open. So almost any day we would have a rally in the last 30 minutes and a sell off in the first half an hour. Do you see such patterns?
This would make sense since investors have essentially been compensated for holding risk (overnight), you are always going to underperform the market if you only daytrade long the indices.