Trade Ideas Blog

This Test Measures Trading Signals: S.M.A.R.T.

Nov 8, 2007

We don’t create alerts unless we know they’re helpful at spotting opportunity for traders. In fact we are now known for removing alerts when market rules change and render some events insignificant.

You can rely on us to keep an ever evolving list of relevant, useful, S.M.A.R.T. alerts.

I explain what that means by describing why we are removing one of our alerts: The Up Tick rule

Is it really doing what we thought?

When we write a formula for an alert or filter we sometimes wonder, “Is it really doing what we thought?” The market makers, specialists, and other traders don’t always show us their real intentions, so we often have to rely on common sense and good guesses.

This is really cool. Now we have proof.

We have an alert called stepping down. This alert looks for the case where someone tries to send a market short order for a large number of shares. Because they had to wait for an up tick, they couldn’t do it in the obvious way. Instead, they had to offer the stock at one penny above the price of the last print. Each time there was a new print for a lower price, they could lower their offer. We recognized that pattern, developed the logic, and displayed a stepping down alert.

Recently the up tick rule went away. Now people don’t have to use that trick. And, sure enough, we don’t see as many stepping down alerts.

(If you’d like to read about other effects the removal of the Up Tick rule has had on the markets, read here (thanks Mike!))

We still see a few. But we saw about 100x as many back when the Up Tick rule was still in effect. What does that mean? That means that the alert worked as intended. Approximately 1% of these alerts were just noise. The other 99% were legitimate patterns.

What it takes to join the alert list

Over the years a few people have asked why we didn’t have a stepping up alert. It’s the same answer. The (legitimate) stepping down pattern was caused by the Up Tick rule. That only applied to short sales. There was no reverse pattern for people buying stock. We could have entered the opposite formula into our servers, but it only would have reported noise. There was no legitimate pattern to report.

We don’t create alerts like that.

S.M.A.R.T. Alerts

So our test for an alert to join the club is: S.M.A.R.T.

  • Specific – Does the alert describe a legitimate, actionable pattern?
  • Measureable – Can the event be quantified or is it subjective?
  • Actionable – Can traders reap rewards from finding this pattern in the market?
  • Relevant – How much noise or false signals might this event generate?
  • Timely – Can the event be spotted in real-time?

All of our alerts pass this test when properly combined with our filters. Use this helpful test in your own trading. Do your signals pass this test? How much noise do you sift through before deciding on a trade? Answers to these questions will tighten the belt of any trading plan and hopefully generate additional insights into your trading.