A fifty-day average breakout as a trend permission filter

A permission line, then a later break

This archive case follows the S&P 500 in January 2005 around a 50-day moving average. Here that moving average is a 50-day average of closing prices, a slowly updating reference line against which daily closes can be judged above or below.

The same line was used as a trend filter: a binary permission rule that allows only long setups while the index is above the 50-day average and only short setups while it is below.

A hover that was not a settled break

From January 5, 2005, the S&P 500 spent several sessions only slightly above its 50-day moving average before any later close resolved that hover.

On January 11, 2005, the index traded through the 50-day moving average during the session and closed just below it, still too close to treat the break as settled.

On January 12, 2005, the index fell well below the 50-day moving average, then closed back above it by only a small margin.

Editorial: While the daily close keeps hugging the reference line, the trend filter has not yet granted a new side. The hover is information about indecision, not a completed breakout.

The close that left the band

A following session produced a last-hour decline that closed well below the 50-day moving average, after which that average acted as resistance.

On January 14, 2005, the index remained below the 50-day moving average even though the close was near that day's high.

In this case a breakout is a later session in which price leaves the moving-average band that had contained it and then either holds that side or returns to retest the same level on a shorter chart.

A shorter-chart retest of 1190

After an upward bias on January 17, 2005, the index broke back above its 50-day moving average near 1190 on January 18.

On a five-minute e-mini chart, price pulled back to the 1190 breakout area around 8:05 a.m. and then continued higher.

Editorial: The January 18 close-side break supplied the new permission to look only long. The five-minute return to 1190 is the second clock, the retest that treats the break as a hypothesis instead of a finished signal.

Volume into the close

Session volume was observed to surge, more often than not, during the final half hour of trading.

Editorial: That volume habit is why a last-hour close well below the average was read as a more decisive leave of the band than the earlier sessions that finished only a small distance from the line.