Anyone doing any historical back testing?

digitalnomad said:
Sure. As much as I can. While I actually trade monthly, what I perceive to be the most optimal inputs for my algorithms, using my pricing model, I simultaneously run thousands of trade simulations across multiple instruments and time series. I choose the most optimal variation at discretion to trade live for the following month. Most would assume this by what I just typed, but the simulations take hours to run, and run on multiple machines :)
More...

Thanks digitalnomad,

I am beginner developer of system trading so my questions may be elementary.

So basically optimize your user input monthly and use them going forward for the next month. I assume you do this to keep your systems parameters up-to-date with monthly market conditions. Which makes sense and something I have been considering rather then testing a intraday strategies on 10 years of historical data. Just my thoughts.

Let me know if my understanding matches what your statements. Thanks
 
Exactly. Ranges and volatility change, and I assess every month. The variations I choose many times are identical month to month, and if not, they don’t drastically differ. I don’t have a need to store over six months historical data, and only collect real-time data
 
digitalnomad said:
Exactly. Ranges and volatility change, and I assess every month. The variations I choose many times are identical month to month, and if not, they don’t drastically differ. I don’t have a need to store over six months historical data, and only collect real-time data
More...
Thank you for response digitalnomad.

So the traditional back test for +10 years of historical data does not apply to your trading system development and real time trading? I agree with this because for me, I trade intraday and normally averages 5-10 trades per week. So after 6 months I can have statically significance on recent market conditions.

And its even better because your are collecting and saving real time tick data day-to-day for free. Then use that data for optimizing for 30 days. Smart idea.
 
Yes. That’s the idea. Take long-term trend following, or even swing trading for example. By the time you gather enough samples to build statistical confidence, you’ll be taking a dirt nap.

The goal of trading IMO is not to grind away at it your entire life, but to exploit anomalies, and leverage them until the music stops. Then reallocate the profits away into lower risk assets (physical preferably), and never look back. IOW you should only be in this to get rich. This business is like a giant game of Russian roulette, especially on the retail side.
 
digitalnomad said:
Yes. That’s the idea. Take long-term trend following, or even swing trading for example. By the time you gather enough samples to build statistical confidence, you’ll be taking a dirt nap.

The goal of trading IMO is not to grind away at it your entire life, but to exploit anomalies, and leverage them until the music stops. Then reallocate the profits away into lower risk assets (physical preferably), and never look back. IOW you should only be in this to get rich. This business is like a giant game of Russian roulette, especially on the retail side.
More...

Hello digitalnomad,

I agree with everything you said!!

Finally someone on ET who thinks like me. I do not like long term trend following or swing trading, that is what S&P500 index for.
 
You must log in or register to reply here.
Back
Top