Does the trading rule mrinasset160 have a negative sharp?

[TobiasAntiGravity]

Hi

First off - thank you Mr. Carver for your books and repo! It is a whole new world of exiting new stuff to digg into and try to understand!

I am assuming that I am at fault here, but not sure how, and so wanted to pose the question just to make sure;

I loaded the provided/rob_system into jupyter, and fed it the csv sim data from the repo. Ran the following command

system.accounts.pandl_for_trading_rule("mrinasset160").percent.cumsum().plot()

which showed the following, steadily declining graph;

system.accounts.pandl_for_trading_rule("relmomentum10").percent.stats()

gave

Looks to me like the system uses the rule as it has forecast weights for a number of instruments.

Disclaimer - not all instruments used by the system had sim data - had to add the instruments with missing data to the exclude_instrument_lists. But, only the following instruments where missing data, and had non zero forecast weights for the mrinasset160 rule;

['EU-INSURE',
'EU-DIV30',
'EU-TRAVEL',
'EU-BASIC',
'EU-OIL',
'EURO600',
'EU-RETAIL',
'EU-AUTO',
'EU-HEALTH',
'EU-UTILS',
'EU-CHEM',
'EU-FOOD',
'EU-TECH']

Seems unlikely that these are essential for a positive sharp - maybe I am wrong?

[robcarver17]

You're not at fault. The trading rule is indeed a money loser in backtest. You could have confirmed that from this blog post

"Trading rule performance
Here are the crude Sharpe Ratios for each trading rule:

breakout10 -1.19
....
mrinasset160 -0.63"

And as I also say in that blog post

"In bold are the rules that are genuine money losers:

• mean reversion in the wings
• mean reversion across assset classes

What these all have in common is they are non trendy, mean reverting, and hence highly diversifying rules. I haven't dropped them, because a proper handcrafting process would give them a positive weight: they are strongly negatively correlated to the trendy rules, and whilst their negative Sharpe Ratio tilts their allocation down a little, the uncertainty about backtested Sharpes means they are still justified a positive weighting."

[TobiasAntiGravity]

Thank you for your for the explaination! This (of course) makes sense!!

Apologies for not searching your blog before posting this! I was sure that I had read in Systematic Trading, that negative sharp rules was not to be used. I was obviously wrong.