Back testers

[hotwired]

Hi all
I'm looking for a good back tester to get myself clear on how various portfolios perform. It doesn't HAVE to have a "trading" feature (i.e. what if I pull out anytime investment falls below 300 EMA, etc.) but it would be great if it did. What I'm looking to discover is how much difference various portfolio allocations make, both complex and simple. This will help me decide on how much to simplify. Now I've got about 18 funds and I find this is too many for me to be "clear" on. However, I'd like to make sure I don't shoot myself in the foot in my simplification. I think a backtester, while not perfect, will at least show me some hints.

[galeno]

GIGO?

I can make a cheap simple globally diversified portfolio with 1 world equity fund/ETF, 1
core bond fund/ETF, and a money market fund.

I fail to see how looking backwards will help you simplify your portfolio. A far better tool would be Morningstar's Instant Xray tool.

I find the Stock Intersection part of the tool to be particularly valuable.

[hotwired]

I do understand that, and I could do the same. However, I have nagging doubts that I want to put to rest. For example, what if I find that I could've earned 2% more per year if I'd just put 12% vs 5% in RE or Gold? Or what if I'd gone to cash when any asset breached the 300 day exponential moving average.

I do also know that even the most impeccable back tester cannot predict the future, but I think I can get a good idea of how simple I can get...

[haha]

Quote:

Originally Posted by hotwired I do understand that, and I could do the same. However, I have nagging doubts that I want to put to rest. For example, what if I find that I could've earned 2% more per year if I'd just put 12% vs 5% in RE or Gold? Or what if I'd gone to cash when any asset breached the 300 day exponential moving average. I do also know that even the most impeccable back tester cannot predict the future, but I think I can get a good idea of how simple I can get...

This is called data mining, and it is generally judged to be useless. If it interest you, and you want to do the work on it, you could then forward test for some years anything you think is promising.

Best of luck!

Ha

[galeno]

Another name for data mining is curve fitting. That's why I said GIGO.

[target2019]

Ezbacktest is useful for this purpose.
You might also search bogle heads site for a backtest spreadsheet.

These are free, and oriented towards mutual funds.

[rbmrtn]

+1 on EZbacktest, here's the link Free Download: EzBackTest

Also ETFreplay for ETF portfolio, Backtest ETF Strategies , some parts are free.

[photoguy]

In general, back testing, curve fitting, evaluating on historical data, is quite useful but can also be very dangerous. Throwing a bunch of s*** against the backtesting wall and seeing what sticks however is not the way to use it.

[SVHoper]

Hi,

Simba's backtesting spreadsheet - Bogleheads

This one will work for asset allocation back testing but not for trading (I don't think it has sufficient time detail.

I have not used this link but I know the spreadsheet is Excel... not Google docs - so you need to download the Excel.

[hotwired]

I am looking to simply eliminate any possible "great strategies" in AA that I might overlook doing a simple 2,3 or 4 way split. I've mostly decided to do a 60-40 split with a SC / Value tilt, going evenly between US and INT'l -- In giving up "timing" or "tactical asset allocation" I still want to be able to rebalance on a more micro level (i.e. small company, etc.) and I want to be at least a little tilted toward the strategies I believe give the best overall long term return (SC / Value). I found the spreadsheet above very useful for this, and thank you!