QuantConnect Community

Not such an optimistic picture when backtesting the strategy over the past 5+ years.

Vladimir Yeah, it does not perform well against the past 5+ years, likely because the equities market has been very hot lately. However, equities may not always be hot and this Asset Allocation Strategy might help generate steady returns in other market conditions. This strategy beats the index when you backtest from 2008 because we are able to observe different market conditions. 

My goal for Adaptive Asset Allocation is to reduce drawdown and generate slow+steady returns. Seems like this can protect you from the shakiness of the market by diversifying into various assets other than equities. 

The returns are less than benchmark, but more stable. My next goal is to increase the returns. 

I have a few ideas and am testing a few things on Portfolio visualizer message me got something to about 18% CAGR and 6% Drawdowns let's connect if you want to collaborate

how can we add stop loss orders to this, and multiple lookback periods would probably help performance a bit or at least lower drawdowns. 

Dual Momentum is useless honestly the long term backtests hit like 40%+ drawdowns which no one shows because they using awesome things like Monthly DD lol.

Hi Jack Pizza

Stop-loss can be implemented by adding the below line in Initialize:

self.AddRiskManagement(MaximumDrawdownPercentPerSecurity(0.02))   # 2% stop-loss

Best
Louis

superb thank you

Also Louis Szeto I'm noticing this algorithm is being a little weird as opposed to the traditional schedule rebalancing, how can we change it to weekly instead of monthly?

nevermind seemed to have fixed it by adjusting schedule on at top and commenting out some of the last month stuff.

Louis Szeto doesn't that create a trailing stop? How do we do a regular stop?

Hi Jack

That was a regular stop, we also have a trailing stop risk management model. Check the docs and source code of the TrailingStopRiskManagementModel.

Best
Louis

Louis Szeto 

 MaximumDrawdownPercentPerSecurity(0.02) is still a trailing stop in monitors the securities unrealized profit so it updates…?

pnl = security.Holdings.UnrealizedProfitPercent

if pnl < self.maximumDrawdownPercent:

it's basically a trailing stop at the security level instead of the full portfolio…. isn't it? 

Hi Jack

Nope. It is an absolute stop. security.Holdings.UnrealizedProfitPercent refers to the percentage change of an asset's opening positions based on the average opening price. 

But if you keep diluting your opening positions, this stop level would be unstable. In this situation, you'll need to place active orders to control each opened position.

Best
Louis

hello everyone, in 2022 the strategy also suffered a hard drawdown. Have there been any developments?