Back to Community
101 Trading Ideas / Silver and Gold

Here is a small variation from one of Henry Carstens "rough" 101 Trading Ideas systems from http://henrycarstens.com/101-trading-ideas-silver-and-gold-gold-silver-strategy-101tradingideas/

This version uses ETFs instead of futures.

This is long-only variation and because "17 days" of holding felt too overoptimized I changed it to round 20.

I usually don't like short/long versions of the same algo (long/short volatility is usually not symmetrical FX being the main exception to the rule)

It is also said on the end of the post that "this also works with crude" so I combined both crude (USO in this case) and gold (GLD) systems. Both have leverage level of 0.5.

Clone Algorithm
91
Loading...
Backtest from to with initial capital
Total Returns
--
Alpha
--
Beta
--
Sharpe
--
Sortino
--
Max Drawdown
--
Benchmark Returns
--
Volatility
--
Returns 1 Month 3 Month 6 Month 12 Month
Alpha 1 Month 3 Month 6 Month 12 Month
Beta 1 Month 3 Month 6 Month 12 Month
Sharpe 1 Month 3 Month 6 Month 12 Month
Sortino 1 Month 3 Month 6 Month 12 Month
Volatility 1 Month 3 Month 6 Month 12 Month
Max Drawdown 1 Month 3 Month 6 Month 12 Month
# Backtest ID: 577cce57d540cb10a29fde79
There was a runtime error.
8 responses

I don't know whether you have had a chance to search some of the freely available and free databases out theresuchas the FRED or the B of E but you will find spot gold prices going back in some cases to the 1850s. likewise silver. Albeit monthly. I believe such deep history would provide you with a more valid view of you strategy.

Anthony, this is just a Quantopian version of the strategy that author described as a rough draft so I'm sure it's not perfect if that's your point?

If you have modified the algo to monthly mode and tested it with mentioned free data from 1850's feel free to post the results. However I wouldn't personally trust very old historical results of algos at all as there have been some very basic strategies that worked quite well until computers became widely used and then have stopped working competely but that's just me and this of course depends on the strategy. Some momentum strategies seem to be very consistent (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2292544)

I posted this because I though some people might not want to implement it themselves (author posted this as a link in trading ideas thread).

I personally think this strategy is quite neat as it's long only and still quite consistent even when uso and gld have not been especially stable in their trend.

Thanks for sharing. I'm a bit surprised that this algo works so well.
I'm unfamiliar with commodity trading, but is the method of 20 day average compared to 20 days prior a common method in commodity trading?

No, I just wanted to say how illuminating I have found such data applied to my own systems.

Far from being outdated or useless I have found it has validated my thoughts and been hugely useful.

Here is another one of Carstens "rough" strategies. Doesn't seem to work as well as it does with futures backtest at his site with etfs (or then there is some problem with my logic or it might be that the DX contract does not correlate with dollar etfs for some reason).

http://henrycarstens.com/101-trading-ideas-effectiveness-strategy-dollar-101tradingideas/

Clone Algorithm
2
Loading...
Backtest from to with initial capital
Total Returns
--
Alpha
--
Beta
--
Sharpe
--
Sortino
--
Max Drawdown
--
Benchmark Returns
--
Volatility
--
Returns 1 Month 3 Month 6 Month 12 Month
Alpha 1 Month 3 Month 6 Month 12 Month
Beta 1 Month 3 Month 6 Month 12 Month
Sharpe 1 Month 3 Month 6 Month 12 Month
Sortino 1 Month 3 Month 6 Month 12 Month
Volatility 1 Month 3 Month 6 Month 12 Month
Max Drawdown 1 Month 3 Month 6 Month 12 Month
# Backtest ID: 578015eb03dabe0f974c151c
There was a runtime error.

I have often wondered whether all these guys make money and how. It seems such a struggle: Covel, Carstens, Nick Radge. Some of them have good ideas, some don't. Most don't provide a real time track record.

Or Ed Seykota - charming and intelligent but what a struggle peddling conferences and seminars.

The best money I have made trading has come from flipping IPOs which is or was virtually risk free, never a losing month etc. Everything else is disappointing by comparison. Hence my ignorant fascination with other schemes such as HFT where the die are also loaded in your favour.

I'm unfamiliar with commodity trading, but is the method of 20 day average compared to 20 days prior a common method in commodity trading?

I think it's just a crude way of checking if "trend" (whatever that really means) is up or down. I did try a different variation of this just for fun and for verification to see if the numbers are overfitted etc and basically changed all the means in the algo to regression slopes. The curve looked quite similar which implies that it's not at least totally curve fitted (and there actually is quite good economic reasoning behind this but I also understand that doesn't matter that much in trading..).