Post New

I noticed lately that after a few days where nothing happened, stock-gambling euphoria reduced a little on forums I read.

On days the market dumps a bit, it reduced more.

On steep down days, it reduces fast.

If anyone is interested in the 99-day 'dying bullish euphoria' rule/MI approach, and if you have a coded-up implementation of the 99-day rule which can be adapted, could you please consider trying this modification and let me know the result?

- Day with no new high, flat or rise: +1 days to clock.

- Day with drop >0.5%: +1.5 days to clock.

- Day with drop >1%: +2 days to clock.

I would love to know if it would be either more predictive or if it would help the spread of results (60-120 day rules) to clump together more.

I don't know if GTR1's capabilities extend to dynamically adjusting signal states based on results within a backtest period. Not answering the call to more analysis directly, but here are some resources from the archives / FAQ / datahelper on the 99day rule about the extensive drlll-around tests that were done historically.

http://www.datahelper.com/mi/search.phtml?nofool=y... - summary of effects of combinations of the bear catchers on common metrics (Klouche).

http://www.datahelper.com/mi/search.phtml?nofool=y... - BarryDTO's work on parameterizing BCIII/DBE and others against different universes. Whole thread view: http://www.datahelper.com/mi/search.phtml?thread=3...

It's easy enough to download the S&P or any index data for several years to a CSV, upload to Sheets/Excel, calculate the 99day signal standard in 1 column, then next to it add those conditions in calc/formula columns. Those adjustments are interesting to think about; at the risk of added complexity.

The prompt is, would those parameters get you in significantly earlier in a rising bull that lasts, or get you out significantly earlier in a bear that does some real damage?

RayVT's post here on the value of timing and pitfall (mainly, investor psychology and the will to act when triggered) was as on point 9 years ago as it is now: http://www.datahelper.com/mi/search.phtml?nofool=y...

I've also dug up some links to original research on using the signals together, and updated the Timing Methods page in the FAQ:

Putting it Together: Using key signals to manage rebalancing & allocations
Breadth combined with short term MACDs/PPOs: http://www.datahelper.com/mi/search.phtml?nofool=y...
Using Breadth Extremes ($NAHL) to identify likely short term bottoms: http://www.datahelper.com/mi/search.phtml?nofool=y...
Identifying Market Tops and Bottoms: http://www.datahelper.com/mi/search.phtml?nofool=y...
In the 2008 GFC, how the Methods worked together to get out of equities before the big damage was done: http://www.datahelper.com/mi/search.phtml?nofool=y...

Cesar Alvarez did backtests showing requiring SPY or QQQ to be above MA200 for multiple days instead of 1 is much better at risk reduction and sometimes return as well.

In my looking at spy and qqq the old Golden Cross 50 day sma above or below 200 sma seems to do a good job. Any ideas about addtional criteria to the 50 and 200 to even make it better?

additional info on 99 day rule high from Allocate Smartly:

A Key New Momentum Measure to Consider: Distance from 1-Year High
This research was inspired by Alpha Architect's coverage of a new paper looking at how the distance from a stock's 1-year high has affected the performance of momentum strategies and the likelihood of 'momentum crashes'. We look at the same question applied to a stock index: the S&P 500. We show that how far the [']

FILED UNDER: TAA ANALYSIS, TAA STRATEGIES

https://allocatesmartly.com/a-key-new-momentum-mea...

Post New

Mechanical Investing FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.