Market regime filter

[Bigchazza]

Hi all. Do you guys put a regime filter in your algorithms? I looked into this, with the idea being to avoid the damage of bear markets. The most common one I see is "if the SPY price < its 200 day moving average, bear market. No trading!" But when I run it in backtest it peforms poorly compared to not having any regime filter. So I am thinking, is it worth it? I get that it is primarily intended to stop me losing everything in a 2008, but it seems to be at the expense of so much lost opportunity.

Do others have a more subtle approach than the above? I was wondering about a risk-based approach i.e. use SPY price against 200 day MA to calibrate how much risk I apply to trades. Would welcome thoughts.

 

[GaryBtrader]

It depends on the strategy. I do use a filter similar to this in one of my strategies and it works well. Overall growth is sacrificed a little bit but you'll be thankful once you trade through a scenario like 2008 and see the losses that were avoided. It's those difficult periods that take an emotional toll on a trader and cause costly deviations from a tested strategy.

 

[MarkBrown]

if you have not such your doomed. so yea that's kinda a big one unless you have basic old fashioned trading systems which ignore regime change.

ps study instead of averages - volatility

 

[Bigchazza]

if you have not such your doomed. so yea that's kinda a big one unless you have basic old fashioned trading systems which ignore regime change.

ps study instead of averages - volatility

Hi Mark Brown. Re your volatility comment, do you mean perhaps a look at the volitility of the index might be a better regime check than average price?

 

[longandshort]

Hi Mark Brown. Re your volatility comment, do you mean perhaps a look at the volitility of the index might be a better regime check than average price?

Yes, volatility tells you how stable the trend is. Markets act as discount mechanisms, so forward volatility is more useful than historical (e.g. VIX). Analyzing volatility, you'll notice clusters (periods of high volatility, periods of low volatility), hence regime instead of mean reversion. You can then learn what optimal strategies would existing in different regimes, and at crosses. Have fun.

 

[MarkBrown]

Hi Mark Brown. Re your volatility comment, do you mean perhaps a look at the volitility of the index might be a better regime check than average price?

yes i have tested both and vol wins.