Last Updated on January 16, 2022 by Oddmund Groette
I have been investing and trading for about 16 years, but not until now have I had a look at pivot points in trading. We can read a lot about pivot points in certain media magazines and by stock commentators.
I am reading Briefing In Play before the US open, and they have a dedicated website to pivot points (but I have never looked at it…). Before writing this article I must admit I had basically no clue what this really is. Some weeks ago I read The Day Trader – From The Pit To The PC by Lewis Borsellino and he wrote how he used pivot points before trading every day (by the way, this is a horrible book). Borsellino was among the best and biggest pit traders in the S&P 500 in the 80s and 90s. When he made so much money, pivot points must be a very good indicator?
I first did a search on the internet and found the formula:
Where C is yesterday’s closing price, H is yesterday’s high and L is yesterday’s low, the central pivot point for today and its support and resistance levels are defined as:
Central pivot point P = (H + L + C) / 3
First resistance R1 = (2*P) – L
First support S1 = (2*P) – H
Second resistance R2 = P + (R1 – S1)
Second support S2 = P – (R1 – S1)
However, after searching the web there seem to be several ways to calculate pivot points. But I will stick to my definition above.
Before I do any testing I would say this does not make very much sense to me. But hey, I am open to new ideas so let us test some ideas based on pivot points.
Now, there are probably a zillion ways to play this. Unfortunately, Borsellino did not explain what he really did, he just wrote it was very useful for him. I will now just test some random strategies based on the following by using the first support and resistance only (pivot point is PP):
- The market opens above yesterday’s close and drops to the first PP support. Entry is first support PP and exit is on the close. Of course, this is a long-only strategy.
- The market opens below yesterday’s close, but above the first PP support, and drops to the first PP support. Entry is first support PP and exit is on the close.
- The market opens below yesterday’s close and rises to the first PP resistance. Entry is first resistance PP and exit is on the close. Of course, this is a short strategy.
- The market opens above yesterday’s close, but below the first PP resistance, and rises to the first PP resistance. Entry is first resistance PP and exit is on the close.
- The market opens below first support. First PP support now becomes resistance and first PP support is now an entry for short. The exit is as usual on close.
- The market opens above the first PP resistance. First PP resistance now becomes support and first PP resistance is now long entry. Exit on close.
As you can see, this is just day trades. As mentioned, the are infinite ways to trade this and I only looked at the first support and resistance. This took a lot more computer power than expected so I did it as easily as possible. These are just some strategies meant for a brief indication and perhaps to give some further ideas.
I have tested these 6 strategies on a basket of 73 ETFs listed on NYSE and NASDAQ. The test period is from the year 2000 until yesterday. Slippage and commission are not included. However, there should at least be no slippage on entry with this strategy.
Here are the results:
|Strategy||# Fills||Wins||Avg.win/avg.loss||Average per trade|
As you can see, this is not very impressive. Short is a total disaster, but that is not unexpected. It is a lot more difficult to make money on the short side (read here for why short selling is difficult). The best strategy, number 1, has this profit diagram ranked on each ETF:
I am not going to go into depth about this, but there are certain patterns on different ETFs that seem distinct. But I doubt that pivot points have any predictive power at all. The reason strategy 1 is profitable is probably due to the upward bias in the markets. However, I got some ideas for further testing. Certain ETFs seem to work much better than others, which I do not believe is attributable to chance and/or curve fitting.