The distribution day count is at 4 days- 10/19, 11/9, 11/11, and 11/12. 10/19 will drop off the count starting on Tuesday, so if Tuesday is a distribution day, the count would still be at 4.
That said, three of the four distribution days occurred over four days, thus are weighted more heavily than had they occurred in a more spread out manner. The action of leading stocks is also taken into account. Some leading stocks held up okay last week but others got hurt. For example, NFLX, FFIV, APKT, IGTE held up okay, but high volume selling occurred on AMZN and AAPL, RVBD had a reversal, and NTAP had a mini-gap down. We will watch for further evidence of breakdown in leaders in the days ahead that could create a neutral or even a sell signal in the model ahead of a break below the 5th distribution day which would trigger a sell signal, and let members know if and when the model switches. Note, keep your sell stops (mental or actual) in place regardless of the model's signal as the model is designed for investment in ETFs. Your strongest stocks may buck a market pullback or a correction so dont sell prematurely just because the model issues a sell signal. Also, investors with longer time horizons who are sitting on profits and/or who prefer to use the 50dma could potentially sit through a market correction without their stocks triggering sell alerts. Some conservative investors may elect to tighten their stops as this may fit their risk tolerance levels, which of course is unique to each investor.
Last week, "stuff" stocks staged major reversals, largely due to fears that China will raise interest rates again. But keep in mind that gold ETF GLD and silver ETF SLV staged major reversals on 10/7 then both turned around and went higher showing the power of QE2 and the belief that the dollar will, overall, continue its long term downtrend. That said, there may be brief periods of decoupling as the correlation between the dollar and the stock market is far from perfect. Also note that China raised interest rates on 10/19, but that didnt stop the US general markets from pausing briefly then continuing its uptrend. Of course, price/volume action of the major indices and leading stocks rule the day and are the final judge as all the news and speculation out there is noise that can derail one's focus from the facts that are price and volume.
Distribution Day Learning Corner
Note, 11/12 is a distribution day despite the lower volume because 11/11 was exaggerated due to S&P index rebalancing and due to CSCO trading an abnormally large number of shares due to its earnings report.
Also note, 10/29 is not a distribution day. It was a very close call. Once in a great while, you get days such as this one that are unusual occurrences that I call 'gray' days. Fortunately, such days are very rare. I studied the effects of such days on the model by seeing the differences between labeling such days as distribution vs. non-distribution days. I found it had little effect on the overall performance of the model first because such days are very rare and second, even if the model issued a sell signal prematurely, had this signal been false, this implied a strong market, so the model would issue a follow through day shortly after thus not miss much of the move higher. Likewise, if the model issued a sell signal late by getting to its fifth distribution day late due to excluding an earlier distribution day, were the market weak, it would still catch the majority of the market's downtrend.
For consistency, such gray days obey the rules I developed thus they arent 'gray' in a rules sense. However, it is always a good exercise to create 'what-if' scenarios to see how the model would be affected had the call been made the other way.