The Best Time To Trade The Fidelity Select Indicator

--------------------------------------------------------------------------------------------------------------------

One of the new charts we've started posting to the site is the percentage of Fidelity Sector funds that have out-performed cash over the past three months.

I've received quite a few questions about how to interpret this. Is it it best used as a confirmation of a strong market, and we're most likely to keep going higher? Or is a signal of an overbought market, and we're most likely going to fall back?

Typically, it means the former. It has been unusual for the indicator to reach 100%, with only 128 days hitting that level since May 1987. The vast majority of those days (111 of them) occurred during 2003 and 2009, the strongest parts of the bull markets.

Buying the S&P 500 anytime Fidelity Select breadth hit 100% and selling three months later would have given you 94% winning trades (120 out of 128 days), with an average return of +5.4%, an average maximum gain of +7.2% and average worst loss of -3.4%. So very, very good.

Outside of 2003 and 2009, you would have still had 15 winning days out of 17 samples, but the average return dropped to +2.0%, with an average max gain of +3.5% and average worst loss of -2.2%. Still good, just not quite as much.

Let's look at how the S&P 500 performed the day following various levels in Fidelity Select breadth. The table below shows the net points gained or lost, along with the average return and percentage of time positive.

Level	Points	Avg Return	% Positive	Level	Points	Avg Return	% Positive
=0	570	0.20%	58%
<10	803	0.07%	55%	>10	38	0.02%	53%
<20	668	0.06%	55%	>20	174	0.02%	53%
<30	872	0.06%	55%	>30	-31	0.01%	52%
<40	929	0.05%	55%	>40	-88	0.01%	52%
<50	757	0.04%	54%	>50	43	0.01%	52%
<60	641	0.04%	54%	>60	200	0.02%	52%
<70	552	0.03%	54%	>70	290	0.02%	53%
<80	565	0.03%	53%	>80	276	0.03%	52%
<90	737	0.03%	53%	>90	104	0.03%	52%
				=100	79	0.06%	54%

The best time to buy, at least for a day, was when we had no funds beating the market. Obviously, stocks were terribly oversold at that point. There were 401 days with 0%, and the average return was nearly three times as high as any other level, with a higher probability of success. Those days made up less than 7% of all trading days, but they made up 67% of the buy-and-hold points.

The best bet out of all of them was trading on the long side when less than 40% of funds were beating the market. Trying to be a trend-follower was disastrous. Buying when Fidelity breadth rose above 40%, and then selling when it dropped below, would have given you -76 points, with 41% winning trades.

The best system I could find, after trying hundreds of combinations, was buying when breadth hit 0% and selling when it recovered back to 50%. That would have netted you 1,176 points (versus 853 for buy-and-hold), with a +8.5% average return and 82% winning trades (14 winners out of 17 trades), with a worst drawdown of -253 points (versus -889 points with buy-and-hold). The average winning trade was +11.0%, while the average loser was -3.6%.

By selling when it hit 50%, you would have missed all of those great trending moves like 2009, but it didn't matter (waiting until even more funds beat cash would have been a mistake).

Using a more-volatile index, the Nasdaq 100, you would have netted 2,402 points (versus 1,812 for buy-and-hold), with a worst drawdown of -393 points (versus -3,900 for buy-and-hold). It gave 76% winners with a +14.2% average. When the market is grossly oversold, it usually pays to buy the grossest stocks for a rebound.