Backtested 20 years of different portfolios

[AriaS]

Hi everyone,
I backtested 20 years of a large amount of portfolios by using this website:
https://www.portfoliovisualizer.com/backtest-portfolio
Result: holding 2-6 largest cap stocks (with yearly revision) outperformed S&P500 by far.

 

[xatta]

Thanks for posting the link. It looks interesting, except you can not change the portfolio at all during the backtest period. But after all it is a "portfolio" visualizer and not a "strategy" visualizer.

 

[1nvest]

Thanks for posting the link. It looks interesting, except you can not change the portfolio at all during the backtest period. But after all it is a "portfolio" visualizer and not a "strategy" visualizer.

They do provide a more dynamic allocation, under the term "Tactical Asset Allocation" you can then enter a range of Funds or ETFs and depending on the criteria, will choose the most appropriate ETF. for example the momentum rotation model, will rotate in and out of the ETFs based on their momentum ranking, so it becomes both dynamic and is a strategy in its own right

This is just their example, you can do much better

 

[TWI]

Thanks for posting the link, will have a look around. I have used https://allocatesmartly.com/ for a number of years which offers similar TAA backtests and real time allocations. I pay $300/yr as I do actually invest capital in these systems and it saves me some significant work on diversification and historcials.
After living these portfolios for 8 years my advice is to consider drawdowns and volatility as the key metrics because you have to live with the returns daily, weekly, monthly and 25% draws are very psychologically painful . I consider returns vs. drawdown together with sharpe and sortino. I prefer to leverage up a smooth return than unleverage a volatile one.

 

[brettb]

Also thanks for the link. I don't know if you can rebalance on certain criteria. I'm keen to backtest buying 52 week (or 50 day) lows then rotating out when the stock bounces.

I've built my own backtester from scratch and now have 600 tickers loaded, with data from 1971 to 2024 (not for all stocks though). In theory the results should be believable. However acquiring stock data from 3rd parties is problematic as "dirty" data can really mess up backtests. Also my code might have bugs (lol).

Well if anyone has access to a backtester I'd be interested in this:

• Buy quality stock at 52 week low (dividend payer, long term grower, i.e. MCD, PEP, ELV, O etc. etc.).
• Hodl for 10% - 50% profit (which is best... not sure but a higher CAGR usually comes from lower profit targets).
• If profitable sell and buy something else.
• If no profit then hodl for a year or two, then bail out at breakeven or loss.
• Additional filter: stay in and out of market depending on market index moving average crossovers.

According to my own backtests and Luke L. Wiley in The 52 Week Low Formula it does outperform the S&P 500 with lower drawdowns.