Not yet listed. Not yet listed. Over the test window the family lost after costs, and the quality bar needs a positive result over at least twenty trades.
A strategy that only buys has nothing to do in a falling market. It can sit out, which is the right answer, but sitting out earns nothing, and a year of sitting out is a year. The way to earn from a fall is to sell short: borrow shares, sell them at today's price, and buy them back later at a lower one. The difference is the profit, and it is as real as the profit on a purchase.
In India, shares can only be sold short within the day. Whatever you borrowed and sold in the morning must be bought back before the close. So this strategy is intraday not by preference but by rule, and every position, long or short, is closed by a quarter past three.
The same signal, both ways
It uses the Supertrend line described in the article on the long-only version. When the line flips below the price, the stock has turned up and the strategy buys. When the line flips above, the stock has turned down and the strategy sells short. Either way the ADX has to agree there is a trend worth following, and with no trend it waits. The exit is the next flip, with a stop and a target at the same distances in both directions.
It never flips straight from long to short on one bar. The first flip closes the position, and the next clean signal, if the ADX still agrees, opens the other side as a fresh trade with its own stop.
Twice the opportunity, twice the whipsaw
On a day that trends, the short side is exactly as useful as the long side, and a sharply falling day is as good as a sharply rising one. That is the case for the strategy, and it is true. The case against is the other kind of day, and large stocks have more of those: a flip up, a flip down and a flip up again before lunch costs three stops where the long-only version would have paid one and then waited. Two directions means two sets of false signals, and false signals are what intraday strategies pay for.
In the test, the family lost after costs, for the same reason as its long-only sibling with the second reason added. Trades too small to clear their charges, and choppy days that cost twice. Whatever the short side earned on the falling days it found was not enough to cover what both sides lost on the days that went nowhere.
A falling day is as good as a rising one. A sideways day is twice as bad. Large stocks have more sideways days than either.
The short side's own hazards
A bought share can fall to zero and no further. A shorted share can rise without limit, so a short position's loss has no theoretical ceiling. The stop and the same-day close keep that theoretical, and in a large stock it stays theoretical, but it is the reason shorting is treated with more care than buying. There are also the practical frictions: not every stock can be borrowed on every day, a sharp bounce as other short sellers buy back can take a position through its stop, and a short somehow left open at the close is squared off by the broker with a penalty. The strategy closes its own positions rather than leaving that to anyone.
If you take one thing away
Intraday long/short follows a stock's turns in both directions, which gives it a falling market to work with and twice the false signals. In our test the second outweighed the first, and the family stays unlisted.
Think a stricter filter or wider stop fixes intraday trend, long and short? Open the rule sheet, change the numbers, and run your own test.
Customize Intraday trend, long and short