Has anyone developed a "dynamic" indicator to gauge overbought/oversold levels based on recovery behavior in response to a spike in the market? (RSI isn't contrasting enough to be a strong indicator of this, unfortunately.)

For example, there was a recent sell off of Apple (AAPL) stock when the market (and Apple) peaked, but Apple quickly recovered (somewhat) after the market corrected upward. This tells me Apple hasn't reached its upper bound resistance point yet, so it's safe to buy high.

Yes, external symbols would be needed to contrast stock behavior with market sentiment/corrections.
Yes, some minimal statistical analysis would be needed to judge significance with each correction.

You might want to take a look at Adaptive Lookback which can applied to various indicators. The frequent the market's swings are, the faster the indicator's dynamic period will be.

QUOTE:
take a look at Adaptive Lookback which can applied to various indicators.

Thanks for the Adaptive Lookback link. It may be useful in identifying swing points in a market index (e.g. S&P500) from which the willingness-to-correct for the stock in question must be measured. It's that willingness-to-correct metric that's going to reveal whether that stock is topped out (overbought) or has room to grow (not overbought as yet).

Unfortunately, the "generic" RSI indicator doesn't measure a stock's willingness-to-correct correlated with market swings. If anyone has developed such an indicator that does (even if only on paper), I would be interested.

I'm thinking one could match up a strong willingness-to-correct condition with a buy-on-the-dip strategy when correlated with market swings, but that's another story (future development).