It's Got to End Sometime

Incorporating the element of time into technical analysis has been a perennial problem. A number of attempts have been made, but a satisfactory methodology of doing so in a systematic form has remained elusive.

As part of the research I have been working on for my next book, I stumbled across the idea of trend-failure probability rates. Given my engineering background, I started to think about trends as having a lifecycle, and considered that the lifecycle could be thought of as being no different from a kitchen appliance.

When a microwave comes off the assembly line, it eventually ends up on someone's kitchen counter, where it is plugged in and begins to provide service. Extensive data has been collected on just how long the average microwave should provide service before it begins to fail. The collected data can be plotted, and it forms the mean failure rate for microwave ovens. In other words, a failure-rate probability curve exists that can explain a microwave's lifecycle.

Market trends are no different. Although they don't emanate from an assembly line, they do pop into existence like a microwave. Once they do, they too have a lifecycle -- and, if one can systematically examine the trends of all stocks, sectors and general markets over a large and representative sample, then a trend-failure probability rate curve can be constructed in much the same fashion as for microwaves. If time is abstracted as bars, then trend failure rates can be computed for varying time frames. Furthermore, given that trends have a qualitative aspect, as well (suspect or confirmed), the failure probabilities can be further refined to consider this attribute as well.

Applying this to the current market, here is the short-term chart for the S&P 500. The current trend has been in place for 43 bars and counting. That's a long time, as we all know, and though nobody is computing trend-failure probability rates, common sense says this has to end at some point -- which is why everyone keeps suggesting it.

S&P 500 -- Daily

Source: Investools.com

View Chart »View in New Window »

Based on my research, the failure probability rate for a short-term time frame, where the qualitative aspect is suspect, is 77.64%. Looking at this chart, the only way for the trend to transition on this time frame is for the index to trade back to a prior swing point low.

I have labeled the nearest one, given the current bar count. That low is 1340.80. As it stands today, the S&P 500 would need to retrace to and close below 1340.80 to change the trend and reset the trend-failure probability rates. This can change, however. For example, if two more bars print without seeing a level lower than 1354.92, then that price point will become the new swing point low. A retracement and a close below that price would then be all that's needed to reset the probability failure rate on this time frame.

Thus, as you can see, trend-failure probability rates provide both the element of time and the projected price point for the eventual failure of all trends.

Switching to the intermediate-term trend, you can see that it isn't nearly as stretched though. The current trend-failure probability rate there is currently only a little over 16%.

S&P 500 -- Weekly

Source: Investools.com

View Chart »View in New Window »

Not only is the failure probability rate quite different, but here you can see that the trend can reset by either a rise or a decline. The swing point high is just above the current price point, and all it would take would be a weekly close above that swing point to reset the failure rate count. Conversely, the other alternative would be a trade back to and below the last swing point low.

Finally, on the long-term chart, the same swing point high on the daily chart has also established itself, on this time frame, as being the most likely method with which the trend failure will end up resetting itself. It needs to do so, too, for the failure probability rates is currently sitting at just over 64%.

S&P 500 -- Monthly

Source: Investools.com

View Chart »View in New Window »

The other nuance that one has to consider is just how many bars are likely to print prior to the next trend transition, since that adds to the effective failure probability rate. For example, in the daily chart, it will likely take two or three bars to reach the lower level to reset the trend -- and, as a result, the effective failure probability rate is more like 82% right now.

Automatically computing trend-failure rate probabilities is yet another tool one can use to both assess and move to decrease risk when trading. In the current environment, while a failure on the short-term time frame is likely imminent, the weekly rates are not stretched at all. Thus, on the short-term time frame, the first retracement that resets trend will most likely be bought rather than continually sold.

Looking a bit longer-term, if the monthly price gets above the swing point high in the next few bars, that will once more relieve the pressure for a reset will on that time frame. It would also add credence, once again, to the thought that the first bout of selling could be short-lived.

Thanks for listening and until next time, just keep trading the charts!

At the time of publication, Little was long SDS.