Intraday Service Moving to VolumeAlert.com

Edit: Link to VolumeAlert track record for those that couldn’t view last night’s link.

Last week’s program trading volume executed as principal, for member firms’ own accounts, remained high at 41.5% of total program volume. That’s the eleventh straight week of 40%+ readings, which has only been seen two other times in the last decade… near the bottom in early 2003, and off and on from mid-2004 through the end of 2005. Both instances preceded bullish periods for the market (see long-term chart). As I noted in my January 25th column…

Long-time subscribers know that this [program trading] statistic can be very useful in determining the amount of institutional interest in the market. The 40% level has been a key level throughout the past decade (see long-term chart). Notice that when we see spikes into this range as the market is skidding lower, the selloff typically comes to a halt soon thereafter. In virtually every case, the S&P is trading at a similar level or higher six months later.

The two-week buy setup discussed in last Thursday’s column on spotting runaway moves with the S&P Oscillator went into effect at today’s close. The 90% R-Squared setup discussed Tuesday also remains in effect until April 7th.

Thursday’s underperformance by the bank sector on what was otherwise a solid up day could be viewed as a short-term negative. When the S&P500 has closed up 1% or more and doubled the performance of the BKX, it’s typically reversed course and closed at a lower level within the next couple of days. The last thirty occurrences are listed in the table below. Note that in 24 cases, or 80% of the time, the S&P closed below the setup day’s settlement within two sessions. That’s better than the S&P’s 63% at-any-time odds, although it’s not quite enough of an edge to be added to the board…

SPX +1%, Doubles Performance of BKX
03/26/09… ???
03/04/09… Lower SPX close one session later
02/03/09… Lower SPX close one session later
12/10/08… Lower SPX close one session later
11/21/08… No lower close within two sessions
11/07/08… Lower SPX close one session later
10/30/08… No lower close within two sessions
10/23/08… Lower SPX close one session later
10/20/08… Lower SPX close one session later
10/13/08… Lower SPX close one session later
06/13/08… Lower SPX close two sessions later
03/24/08… Lower SPX close two sessions later
02/13/08… Lower SPX close one session later
01/14/08… Lower SPX close one session later
10/31/07… Lower SPX close one session later
08/24/07… Lower SPX close one session later
08/22/07… Lower SPX close one session later
03/06/07… Lower SPX close one session later
05/25/06… Lower SPX close two sessions later
04/21/05… Lower SPX close one session later
11/12/03… Lower SPX close one session later
05/09/03… No lower close within two sessions
04/10/02… Lower SPX close one session later
10/03/01… Lower SPX close one session later
08/24/01… Lower SPX close one session later
07/25/01… No lower close within two sessions
06/05/01… Lower SPX close one session later
05/21/01… Lower SPX close one session later
04/25/01… No lower close within two sessions
04/19/01… Lower SPX close one session later
04/17/01… No lower close within two sessions

It’s been about a month since I last discussed the intraday Volume Alert service, and this will be the last post on the topic (at Market Tells). The reason is that we’ve created a separate website solely for the intraday service at VolumeAlert.com. The site will be live starting Monday, and all premium subscribers will have access to realtime alerts via email and text message. End-of-day subscribers will also have access to the site for the first month but without email or text message alerts.

I’m very encouraged by some key refinements made to the service earlier this month. The most significant improvement was the inclusion of a key component of the original TrendCatcher strategy – TICKscore. Originally, the Volume Alert service was based solely on volume (see my February 12th column.) Results were good, but I didn’t like the large number of losing signals (47%) and continued to look for ways to improve overall accuracy. The answer lied in combining our intraday volume analysis with the TICKscore indicator. This combination leads to fewer trades that are triggered later in the session, but accuracy is up sharply and the vast majority of trades are held into the close. Here’s a rundown of some key statistics…

Volume Alert performance past 24 months
Overall trades… 166
Winning trades: 112
Percentage Winners: 67%

Overall profit: +71.6% in 24 months

Average profit: 0.71%
Average loss: 0.45%

Largest profit: 4.8%
Largest loss: 1.8%

Trades held into the close: 138
% of trades held into the close: 83%

Peak-to-trough Drawdown: 2.7%
see 24-month equity curve

You can see from the detailed track record that trade frequency is down to an average of 7 trades a month. The reason is that the strategy is more selective in picking entries. It remains 100% mechanical, but it now adapts on the fly for changes in volatility. This means it demands to see even stronger NYSE TICK action and volume during periods of excessive volatility, while it automatically relaxes those requirements when conditions are less volatile. The stop automatically adjusts in a similar manner, although you can see from the trade-by-trade track record that it’s unusual for the stop to be triggered (only 28 out of 166 signals were stopped). I believe this is my best effort yet at attaining the original goal of the TrendCatcher strategy – to hop on an existing trend that’s unlikely to change and ride it into the close.

I’ve already been utilizing this version of the Volume Alert signal since March 10th, so if you’ve detected a slight difference over the past two weeks, now you know why. We’re in the process of moving the Volume Alert service over to its own domain at VolumeAlert.com. This will be complete by the beginning of next week, at which point text message alerts will be reactivated for premium subs (thanks for your patience with this matter).

Categories : Archived EOD Commentary, End-of-Day Commentary

 Print This Post