SPY fell 4.4% today and the VIX jumped to 28.2. In tests back to 1990, a VIX above 20 is when all four CIT dip buys did their best: about twice the edge per trade of calm days.
Today
| Market | Close | Today |
|---|---|---|
| S&P 500 (SPY) | 733.00 | -4.4% |
| Nasdaq 100 (QQQ) | 647.05 | -6.5% |
| Russell 2000 (IWM) | 270.98 | -4.6% |
| Dow (DIA) | 514.27 | -3.5% |
| VIX | 28.18 | +69.7% |
SPY closed above its 200-day average (727), 0.8% over it. That matters in the next table, and it is a thin margin: another day like today would put SPY in the weaker row.
After days like this
Every day since 1993 that SPY fell 3% or more with the VIX above 20, counted once per sell-off (the first day, then nothing for a month), and SPY's gain after it:
| Sell-offs | 1 week | 1 month | 3 months |
|---|---|---|---|
| All (43) | +0.62% 56% up | +1.29% 63% up | +4.93% 77% up |
| SPY above 200-day (16) | +1.99% 81% up | +1.79% 75% up | +6.03% 88% up |
| SPY below 200-day (31) | -0.11% 45% up | +0.56% 55% up | +4.30% 71% up |
| Any day, for scale | +0.23% 59% up | +0.96% 66% up | +2.89% 72% up |
Each cell is SPY's average gain and how often it was higher. The worst 3 months after one of these days: -7.8% when SPY was above its 200-day, -29.2% when it was below.
Sixteen is a small sample, and it holds the worst case too: on February 24, 2020, SPY was also above its 200-day, and a month later it was 24% lower. Below the 200-day, the first two weeks were usually worse before they were better.
Why the VIX matters for the dip buys
The four CIT dip buys buy short, sharp drops and sell on the first strong close, usually within 3 to 5 days. Split by the VIX on the day of the signal, every one of them did better when fear was high. Edge is the trade's gain minus the average of 50 random entries in the same stock held just as long, so it is what the signal added beyond being in the stock.
Edge per trade, 2005 to 2026: Dow 30 stocks, IWM, GLD, TLT
| Rule | VIX 20 or less | VIX above 20 | VIX above 25 |
|---|---|---|---|
| QS RSI | +0.25 | +0.83 | +1.37 |
| CIT MACD Dip | +0.18 | +0.98 | +1.39 |
| CIT ComboRSI Dip | +0.18 | +0.74 | +1.26 |
| CIT RSI+VWAP Dip | +0.09 | +0.45 | +0.76 |
Edge per trade, 2005 to 2026: SPY, QQQ, S&P 500, Nasdaq 100
| Rule | VIX 20 or less | VIX above 20 | VIX above 25 |
|---|---|---|---|
| QS RSI | +0.39 | +1.01 | +1.28 |
| CIT MACD Dip | +0.62 | +1.29 | +1.70 |
| CIT ComboRSI Dip | +0.25 | +0.78 | +1.25 |
| CIT RSI+VWAP Dip | +0.43 | +0.73 | +0.98 |
The same test on 1990 to 2004, an earlier stretch of data, came out the same way: all 8 rule and market pairs had more edge with the VIX above 20 (stocks +0.30 to +0.50 points on calm days, +0.64 to +0.94 above 20). Winners ran roughly 70% for stocks and 70 to 85% for the indexes in every bucket; what changes is how much the winners make.
The price: bigger losers. With the VIX above 20 the worst QS RSI stock trade was -38.9%, against -18.9% on calm days. The rules have no stop; they wait for the bounce, up to 60 days.
The CIT Pullback
The last three months
The other side of the same table: over the last three months the live record of the four rules was 3,705 trades, 57% winners, 0.53 points a trade behind random days. The VIX averaged 16 and closed above 20 on two days. That is the calm-market column, and it is why a quiet market is not the time to judge a dip buy.
Today's list
At today's close 118 of the 509 stocks and funds scanned had at least one dip buy fire (the number in brackets is how many of the four): MPC (4), P (4), PLTR (4), PSX (4), ADI (3), APA (3), CMI (3), COP (3), CVX (3), DELL (3), DOW (3), EOG (3), EXE (3), EXPD (3), GD (3), IEX (3), ILMN (3), INTC (3), KMI (3), LITE (3), LMT (3), LYB (3), MRVL (3), MU (3), NDSN (3), NUE (3), OKE (3), OXY (3), PANW (3), SLB (3), STLD (3), TDY (3), TEL (3), TRGP (3), TSLA (3), TXN (3), URI (3), VLO (3), XOM (3), AEP (2) and 78 more. Among the funds: QQQ, IWM, SPY.
61 earlier trades are still open. The list, with each one's exit level, updates after every close in the CIT Dip Buys watchlist in OddsTrader. A signal is a rule firing, not a recommendation.
Exits: the next open after a close above the prior day's high (the CIT MACD Dip sells at that close); the CIT RSI+VWAP Dip sells the next open after RSI(3) closes above 55. 60 days at most.
How the tests were run: daily bars, fills as stated, no commissions or slippage; the sell-off table uses SPY with dividends. Past results don't promise future ones, and this is not investment advice.
