Replies to post #2777 on Strategy Central

[Walkingshadow]

This is an example of the differences between Stochcharts and AskResearch. In Stockcharts, there is a new stochastic sell signal, and it occurred in relatively overbought territory (I define overbought and oversold on the 15-5-5 stochastic as >80 and <20 respectively, which is pretty standard). But in AskResearch, there is now a new BUY signal as of Friday. Furthermore, that is reinforced by other bullish technical signs (e.g., rising Williams, positive MACD signal, PROC crossover). On the weekly chart, there is a stochastic buy signal as well. So personally, I would not short SPX at this point.

And as you point out, there is some negative divergence, since the most recent peak in the stochastics was lower than the prior peak. However, I do not consider the difference in this case to carry a lot of significance. I would call that modest bearish divergence. Bearish divergence can resolve differently, it does not always mean there will be a big correction. Sometimes there is only a moderate pullback, sometimes none at all.

Many predicted a pullback or even a major correction based upon bearish divergence, and were proven wrong. Some insist that this will still occur, finding all kinds of formations that I consider to be unreliable or just plain bizarre to try to support that interpretation, but I don't buy any of it.

Rather than simply identifying that divergence is present, I think it is important to determine how marked that divergence is, how long it has been developing, and how consistent it is across varying indicators. In the case of SPX, I don't consider this significant.

For example, look at what happened in late December and again in late January in the chart below, and how that resolved. There was bearish divergence, but it resolved with prices moving higher just the same. The divergence in December surprised me, I thought that was pretty significant and that prices would move lower, so I went short. But the break of that red trendline proved that was not going to happen, and I had to scramble to flip long. Frankly, if I had that trade to do over again, I would do it exactly the same way. Probably at least 8 times out of 10 I will be correct.

I think the key with these kinds of things is to remember that you are trading based upon probabilities. You will be wrong no matter how careful your analysis is, and probably you will be wrong about 30% of the time (or more). So... clearly establish what criteria you will use to conclude that your assessment was wrong, and then adjust accordingly when indicated. After reviewing things, if you are satisfied you didn't make some stupid mistake, just quickly conclude that this was one of the "30%" of times you will be wrong, and move on. Worst thing to do is to cling to your initial assessment stubbornly and insist that Mr. Market will see things your way. Well... Mr. Market has never been accused of being reasonable, and is not listening to what you or I insist he ought to be doing, and if you keep insisting you are correct, you will almost invariably watch a small loss turn into a bigger and bigger loss.

So, for SPX I see a sell signal in the Stockcharts stochastic but a buy signal in the AskResearch stochastic (which I consider superior), other positive technical developments, and unimpressive divergence.

[Walkingshadow]

<< The signal seems so quick that it seems to overcome the fact that the Stochastics, which usually fail in a trending market, still give good signals. Would this be an accurate statement? Or, what other things do you look at when the market is trending strongly, which may in turn minimize the effectiveness of the Stochastics? >>


Good point. Stochastic crossovers work best in sine wave sorts of trending markets. The weekly charts tend to give more of this kind of thing; that is, in the intermediate-term and long-term, many markets trend reasonably cyclicly, though that may not be the case on the short-term charts. That's one reason I always look at the weekly charts.

But some will trend strongly for many months from time to time. In that case, the stochastics (like the BBs) tend to have little value. In that situation, I usually just align myself with the trend, and examine things like the ADX more closely, also other technical indicators. And... I try to get a feel for when sentiment is getting too extreme in that particular market or stock.

The ADX in particular tends to give you a good feel for the strength of the trend. Also, the BB width may often be helpful. So if the BBs have been pushed wide open, and the stock or index has just traded right along one rail as it was pushed open, then eventually this becomes as unstable as when the BBs are tightly contracted.

The definition of the extremes of BB width are specific for each stock or index and must be established by looking at the history of the BB width, relative to price movement. In Stockcharts, you can plot the width as one of the indicators.

WS