renko

[born again]

can someone please tell me why i can run a backtest and a run walk forward optimization with ONE parameter not five or six different indicators on a 1,5,15 minute chart and get results that suck, and run the same test using a 10 renko bar and get good results. I can go through the trades looking at the chart after test has been run and i have to say that it does look good, am i missing something here, i really dont see how curve fitting could be a problem. The tests were run on oil (CL). Never really considered trading using renko bars i have to say should i. Thank you for your help

[koganam]

can someone please tell me why i can run a backtest and a run walk forward optimization with ONE parameter not five or six different indicators on a 1,5,15 minute chart and get results that suck, and run the same test using a 10 renko bar and get good results. I can go through the trades looking at the chart after test has been run and i have to say that it does look good, am i missing something here, i really dont see how curve fitting could be a problem. The tests were run on oil (CL). Never really considered trading using renko bars i have to say should i. Thank you for your help

Just as I said to CSharpTrader when he asked the same question, more or less:

Here is the post to which I will refer you: http://www.ninjatrader.com/support/f...65&postcount=1

It might make sense to read the thread from the beginning. http://www.ninjatrader.com/support/f...d.php?p=315433

[Harry]

Renko bars have been created to make look charts good. They do not show real price action. In particular a Renko bar does not show the correct open, high and low of the time period that it covers. This has the following consequences:

(1) All indicators that use high or low of bars, such as current day OHLC and prior day OHLC, will show false values.

(2) A NinjaTrader backtest is run in CalculateOnBarClose = true. If you enter or exit with market orders, NinjaTrader will use the bar open of the bar after the signal bar to determine the entry and exit prices. As we know that Renko bars show false bar opens, all Renko backtests are false.

(3) The only backtestable Renko bar that I have tested is the BetterRenko bar that comes with the BetterBrick Style. They are not as regular as the standard Renko bars, but you can use them for a backtest. They are backtestable because the OHLC of the bar represents the price action.

Below is a comparison of Renko, BetterRenko and UniRenko. If it looks good, it is fake.

[born again]

Harry thank you for clarifying this for me, i have tried to find a downloadable version of the better renko bar with no luck do you know where to find these. Thank You

[Harry]

Harry thank you for clarifying this for me, i have tried to find a downloadable version of the better renko bar with no luck do you know where to find these. Thank You

https://www.bigmiketrading.com/download/vip_elite_circle/593-download.html?view

Membership required which comes at a low one-time fee. You will also find a few hundred indicators for NInjaTrader and about 10 other Renko bar types. I am a simple member of that forum, no financial link to Big Mike, who is running it.

[shanemcdonald]

hi
I have developed a few automated strategies that work quite well on renko bars with crude oil. These same strategies get destroyed when run on most other bars.
I have tested them on pretty much every timeframe and bar type.
I do not backtest at all anymore. All testing is forward on live data.

I do test also on unirenko and better renko. Although the backtesting was more accurate with them, I decided that the live results were more important.

I did not have success with any of the short term scalping strategies on renko bars.
Swing trades on an intrady basis is where I found success.

So, even though you may not be able to backtest with renko bars, you should still consider working with them.

Personally, I consider bar types an indicator, just like any of the other indicators.
My automated strategies work on price movement only. If I were trading in a discretionary manner, I would use time based charts, which show volatility as well as price action.

[Harry]

hi
I have developed a few automated strategies that work quite well on renko bars with crude oil. These same strategies get destroyed when run on most other bars.
I have tested them on pretty much every timeframe and bar type.
I do not backtest at all anymore. All testing is forward on live data.

I do test also on unirenko and better renko. Although the backtesting was more accurate with them, I decided that the live results were more important.

I did not have success with any of the short term scalping strategies on renko bars.
Swing trades on an intrady basis is where I found success.

So, even though you may not be able to backtest with renko bars, you should still consider working with them.

Personally, I consider bar types an indicator, just like any of the other indicators.
My automated strategies work on price movement only. If I were trading in a discretionary manner, I would use time based charts, which show volatility as well as price action.

I fully subscribe to some of your statements:

"Personally, I consider bar types an indicator, just like any of the other indicators."

This is the key to working with Renko or HeikinAshi bars. It is not price what you see, but an indicator.

"I do not backtest at all anymore. All testing is forward on live data... So, even though you may not be able to backtest with renko bars, you should still consider working with them."

Just made this post, because I wanted to repeat what you have written. I fully agree.

[koganam]

... "I do not backtest at all anymore. All testing is forward on live data... So, even though you may not be able to backtest with renko bars, you should still consider working with them."

Just made this post, because I wanted to repeat what you have written. I fully agree.

Not quite sure that I agree fully with that one.

To me, the only purpose of a backtest, is to find out if a strategy would have made money, if we had perfect knowledge of the future. Backtest is the only situation in which we always know what happened after the trade entry. If the backtest cannot make a profit, then for me, there is then no point in testing forward on live data.

IOW, I use backtesting to eliminate those strategies that sounded like a good idea when I first formulated them, but that prove to be a dud, even when I am essentially curve-fitting. In my experience, most strategies fail in backtest, so I get to cut my wasted time really short, and can devote my time to forward testing only those strategies which have shown that they could probably make a profit. I only backtest on bars which show the real price action, mostly Range bars.