11/6, 07:39 AM

While the semiconductor ETF rebounded more than 7% over three days, the trading signal said not to buy to the end, and on Friday it gave back the whole rebound in one day. Was this week's signal right? Now that even the S&P 500 ETF signal has turned off, what should I do?

2026-W44


This week's signal was right twice, and one call is pending

The system did three things this week. It did not enter the semiconductor rebound, it moved Google below its stop line on earnings day as last week's warning said, and it switched off the entry condition for the S&P 500 ETF (SPY) on Friday. The first two were right by outcome. The third cannot be judged yet.

Semiconductors: a buy signal was avoided by one day

The entry condition for thematic ETFs is a MACD crossover. The MACD line, which shows short-term trend, must break above the signal line. The Trail buffer shows, in percentage points of the peak, how far the current price is from the stop line at 80% of the peak.

DateCloseMA10MA50MA150MACDSignal lineTrail bufferFrom peakEntry condition
10-26541.62549.52574.17554.84-14.24-6.910.62%-19.38%Not met
10-27557.98546.50573.45555.93-13.00-8.133.05%-16.95%Not met
10-28567.67546.80573.40557.06-11.11-8.734.50%-15.50%Not met
10-29561.67547.08573.42558.26-9.97-8.983.60%-16.40%Not met
10-30522.49545.35572.62559.25-12.10-9.60-2.23%-22.23%Not met

The gap between MACD and the signal line narrowed from 7.33 on 10-26 to 0.99 on 10-29. One more day of rebound would have been close enough to switch on a buy signal. Instead of that day, the Intel guidance shock arrived and the gap widened again to 2.50. Buying at the rebound peak, the 10-28 close of $567.67, would have meant -7.96% in two days.

The Trail buffer warned before the signal did. It was in the sub-5% risk zone all five days, and even on the widest day, 10-28, it was 4.50%. Being 4.5% above the $537.46 stop line means a single -5% day would trigger liquidation. SMH's daily moves this week ranged from +3.02% to -6.98%, so a buffer like this gave no basis for betting on the rebound. MU, AVGO and BOTZ also briefly recovered their stop lines and then lost them again in the same pattern.

A long-used interpretation holds that when a support level breaks it becomes resistance next time, and that a semiconductor index falling back below its 200-day line within days of crossing above it is a bad sign (StockCharts, 2004-12, 2005-10). SMH recovering its stop line this week without settling above its moving averages fits this pattern. The last line SMH now rests on is the 200-day line at $519.50, 0.6% below the close.

Google: last week's warning arrived exactly

The W43 Q&A calculated that if Alphabet gapped -11.6% on earnings day as in July, it would fall below its stop line, and recommended trimming some ahead of earnings. On 10-29 GOOG fell -12.29% in one day and its Trail buffer dropped from 9.68% to -0.11%. At week-end it is at -1.13%, in the liquidation zone. Because earnings came after the close and the stock opened with a gap the next day, a stop order could not have prevented it. Reducing the position before the announcement was the only defense, and that judgment was right.

SPY: this breach is a 0.04% difference

DateCloseMA10MA50MA150Entry conditionTrail bufferFrom peak
10-26765.37768.06766.72739.88Met27.92%-2.08%
10-27769.66767.22766.70740.68Met28.47%-1.53%
10-28772.31767.49766.83741.47Met28.81%-1.19%
10-29769.36767.25766.87742.32Met28.43%-1.57%
10-30755.72766.50766.77743.16Not met26.69%-3.31%

The 10-day line (766.50) fell below the 50-day line (766.77) by $0.27, or 0.04%. That is smaller than the 0.39% gap between the two lines when QQQ broke down last week. The 50-day line is still above the 150-day line, and the close is 1.7% above the 150-day line. The stop line is far away at $547.13, 70% of the peak of $781.62.

The SPY signal does not mean the trend has broken; it means the reason to buy more has gone. With the two lines nearly touching, one day of rebound next week could switch the signal back on. A signal switched on that way is hard to trust for the same reason. In weeks packed with earnings and data like this one for semiconductors, signals flip day by day.

Position actions

NameStatusActionNext decision criterion
SPYEntry condition lost (0.04% gap), Trail 26.69%Hold, stop addingResume buying if the 10-day line holds above the 50-day line for 2 trading-day closes
SMHBelow stop line, MACD falling againNo new entriesOnly when a recovery of the $537.46 stop line and a MACD crossover occur together
GOOGBelow stop lineExit remaining holdingsNo re-entry before next quarter's capex guidance is confirmed
TSLANew entry 10-26, +7.41% for the weekKeep half sizeWhether the entry condition holds
  • If the SMH 200-day line at $519.50 breaks on a closing basis, defer rebound attempts by a month. AMD earnings on 11-03 are that test. If AMD confirms slowing PC demand as Intel did, this line is likely to break.
  • Do not enter names with a Trail buffer below 5% even if the signal turns on. The cost of waiting for a signal in a rebound is missing the peak, and the cost of not waiting was -7.96% this week. For names near the stop line, the latter cost is far larger.
  • Split new money among names whose entry conditions are intact. Among core indexes, none kept its entry condition. The remaining names are XLE, XLV, AAPL, MSFT, LLY and TSLA. Of these, do not chase this week's surge in XLV and LLY, and judge AAPL after confirming its 11-02 earnings.

No signal debate was held this week. The assessment above is a post-hoc review based only on figures from the signal history. Detailed charts and signals for other names are available on the /signals dashboard.



Related Questions