Would you trust an online backtesting tool to build strategies ?

At the end of the day the only thing that matters
in Trading/investing is if the backtesting earns
profit. I never made a dime from backtesting
strategies and then trying them on a live account.

But forward testing in sim mode prior to live
trading now earns me a decent income.

Yup..it's hard work to forward test vs pushing
a button on backtesting software.
 
mason macgregorson said:
I ran robots form Targets Trading Pro and Tend Hunter for 5 days in real time - paper trading (288 trades) and then ran the same past 5 days with back testing - 37% discrepancies - mostly around entries, targets and stops; 16% missed trades all together and a 23% discrepancy in final P/L. I give Ninja back test a worthless rating. Real time paper trading rules, back testing drools.

What actual side by side data do you have?
More...

It's not your modeling software that is causing that. It's a misunderstanding about what a trade price represents. You don't get to buy or sell at the last trade price. Over a large dataset you will always find the slippage moves against you. For highly liquid stocks the percentage is small. As there is less liquidity the percentage increases. It is important to model and test your order entry algorithms.

The reason these canned back testing software platforms don't do this is because few of their customers appreciate or care about the mechanics of performing a trade. It's also complex and requires a lot more data to do it correctly. Few people have access to all order book changes at all the exchanges, for example. That is an absolutely massive amount of data that dwarfs the size of the tick data.

All my strategies track and attempt to optimize order entries. It is a critical aspect of building a quantitative trading system.
 
mason macgregorson said:
I ran robots form Targets Trading Pro and Tend Hunter for 5 days in real time - paper trading (288 trades) and then ran the same past 5 days with back testing - 37% discrepancies - mostly around entries, targets and stops; 16% missed trades all together and a 23% discrepancy in final P/L.
More...

You are probably also aware that over fitting is your biggest challenge. Ninja Trader is simply doing linear regression. It's like the guy who has a hammer and thinks everything is fixed but hitting it with his hammer. Drawing a straight line through a cloud of data points is important analysis but it's pretty basic. For me it's just one tool in the toolbox.
 
FreeGoldRush said:
It's not your modeling software that is causing that. It's a misunderstanding about what a trade price represents. You don't get to buy or sell at the last trade price. Over a large dataset you will always find the slippage moves against you. For highly liquid stocks the percentage is small. As there is less liquidity the percentage increases. It is important to model and test your order entry algorithms.

The reason these canned back testing software platforms don't do this is because few of their customers appreciate or care about the mechanics of performing a trade. It's also complex and requires a lot more data to do it correctly. Few people have access to all order book changes at all the exchanges, for example. That is an absolutely massive amount of data that dwarfs the size of the tick data.

All my strategies track and attempt to optimize order entries. It is a critical aspect of building a quantitative trading system.
More...

What actual side by side data do you have?
 
FreeGoldRush said:
It's not your modeling software that is causing that. It's a misunderstanding about what a trade price represents. You don't get to buy or sell at the last trade price. Over a large dataset you will always find the slippage moves against you. For highly liquid stocks the percentage is small. As there is less liquidity the percentage increases. It is important to model and test your order entry algorithms.

The reason these canned back testing software platforms don't do this is because few of their customers appreciate or care about the mechanics of performing a trade. It's also complex and requires a lot more data to do it correctly. Few people have access to all order book changes at all the exchanges, for example. That is an absolutely massive amount of data that dwarfs the size of the tick data.

All my strategies track and attempt to optimize order entries. It is a critical aspect of building a quantitative trading system.
More...

not according to Ninja
 
mason macgregorson said:
not according to Ninja
More...

lol. Ninja is not going to tell you how this really works the way I did. It's not because they are hiding something. They just don't know/care because efficient order execution is not what their customers want to pay for.

You are looking for a source that you feel is reliable so you can adopt their beliefs as your own. You will get ahead much faster if you learn the mechanics of what is happening and apply objective thought. Forget what some retailer says. No one gets rich by carefully reading about what someone else is selling.

Also understand that the link you provided at Ninja gives reasonable information. I'm not in disagreement. I was simply explaining to you why those facts exist.
 
Last edited:
FreeGoldRush said:
lol. Ninja is not going to tell you how this really works the way I did. It's not because they are hiding something. They just don't know/care because efficient order execution is not what their customers want to pay for.

You are looking for a source that you feel is reliable so you can adopt their beliefs as your own. You will get ahead much faster if you learn the mechanics of what is happening and apply objective thought. Forget what some retailer says. No one gets rich by carefully reading about what someone else is selling.

Also understand that the link you provided at Ninja gives reasonable information. I'm not in disagreement. I was simply explaining to you why those facts exist.
More...
actually I back test in Trade Station, matches nearly to the penny.
 
Amibroker ----> The world's fastest portfolio backtesting and optimization

Amazing speed comes together with sophisticated features like: advanced position sizing, scoring and ranking, rotational trading, custom metrics, custom backtesters, multiple-currency support


mason macgregorson said:
we are not discussing cost or ease of use but back testing
More...
 
You must log in or register to reply here.
Back
Top