Backtesting collaboration ?

[Partum]

What would be the benefits to help you?

You provide potential strategies, and the other end provide time and tick data?
For what in return?
Where's your sales pitch?

The other trader/backtester would get the strategy. If profitable, a good trade for both parties which would hopefully lead to further collaboration. If unprofitable, then no money lost by either party.

 

[fan27]

The other trader/backtester would get the strategy. If profitable, a good trade for both parties which would hopefully lead to further collaboration. If unprofitable, then no money lost by either party.

There is a major false assumption here. If I spend 40 hours on this project, that is 40 hours I wasn't getting paid for my software consulting services which could equal anywhere from $2,000 to $8,000, depending on how much I am billing my time for. To incur that opportunity cost, you would have to bring something to the table in terms of track record, industry connections or something that would make the situation worth considering.

 

[minmike]

I'm new to EliteTrader so I didn't know there were a lot of requests for assistance with backtesting. Sorry for agitating the curmudgeon crew.

I posted on the assumption that there are more people who know how to backtest than have profitable trading strategies. Ultimately I've ended up getting a few responses of people willing to assist me so lets hope, as you say, they are worth their salt and can operate their backtesting software and we both end up with a profitable system.

Again, thanks for the Ninja suggestion, I'll take a look but what I need tested doesn't use indicators that have variable parameters like MAs and RSI, MACD and the lot, it simply uses price range ratios.

How will you know if they are "worth their salt"? You need way more expertise to check others backtesting than to back test yourself.

 

[minmike]

Why do you believe your strategy is worth testing?

 

[Partum]

There is a major false assumption here. If I spend 40 hours on this project, that is 40 hours I wasn't getting paid for my software consulting services which could equal anywhere from $2,000 to $8,000, depending on how much I am billing my time for. To incur that opportunity cost, you would have to bring something to the table in terms of track record, industry connections or something that would make the situation worth considering.

My assumption is that someone would view this from the retail trading perspective and not the professional software consulting perspective. Some people want to trade for a living or as a hobby and are willing to take some risks with their time and money and some want to program for a living risk free on a contract basis as an employee. I am looking for the former.

Would an assumption that subjective opportunity costs can very quickly become meaningless compared against the discovery of a few profitable strategies be true, strategies that do not use time based bars or indicators that use arbitrary parameters, strategies that will generate trade signals everyday and in all markets?

Many years ago I had a friend who used Tradestation (he has since died). I would give him rulesets over the phone and he would program them in 15 minutes and then let the backtests run. Again, I may be completely ignorant but I am going on the assumption that those who have the inclination to write code and own backtesting software are adept at using it and are always looking for few new trading concepts.

 

[Partum]

Why do you believe your strategy is worth testing?

From preliminary manual backtesting and live sim testing.

Because the strategy relies upon dynamically changing range/tick bar sizes based on volatility it is extremely tedious to manually backtest.

 

[Partum]

How will you know if they are "worth their salt"? You need way more expertise to check others backtesting than to back test yourself.

Timestamps on a portion of backtested trades will suffice. I can manually verify a sample.

I believe the concept to be robust enough that a tick or two variance from someone's tested dataset to my datafeed will not see the strategies imploding. The strategies are not HFT or scalping for ticks.

 

[Slope Trader]

Timestamps on a portion of backtested trades will suffice. I can manually verify a sample.

I believe the concept to be robust enough that a tick or two variance from someone's tested dataset to my datafeed will not see the strategies imploding. The strategies are not HFT or scalping for ticks.

Partum,

I think your idea is a good one. There are lots of people out there who have learned how to backtest but have not been able to develop any consistently profitable strategies.

Good luck to you, hope you find a competent and honest backtester to work with.

 

[minmike]

Timestamps on a portion of backtested trades will suffice. I can manually verify a sample.

I believe the concept to be robust enough that a tick or two variance from someone's tested dataset to my datafeed will not see the strategies imploding. The strategies are not HFT or scalping for ticks.

From preliminary manual backtesting and live sim testing.

Because the strategy relies upon dynamically changing range/tick bar sizes based on volatility it is extremely tedious to manually backtest.

Good luck. Unless "dynamically changing range/tick bars" are part of someones backtesting package that is a real project to create from scratch. As is working with tick data in general. What you are asking is well eyond the capabilities of all but the most extreme outliers of retail traders.