Autochartist

[cassiopeia]

Has anyone recorded any past predictions of chart patterns from this software?

 

[shadowninja]

Tried it for a month - I'm not that impressed. Often the R/R was screwy or it was doing stuff I just wouldn't do. The chances of it being right are also not that great. I don't think you can just use it blindly - you will have to perhaps only take trades in the direction of the overall sentiment... or maybe flip a coin.

 

[lamog7]

Chart patterns have 50/50 win rate over the longer term. Forget about them.

 

[forker]

That depends on your definition of chart patterns

 

[shadowninja]

That depends on your definition of chart patterns

No, it doesn't. It's the chart patterns that are specifically identified by Autochartist.

 

[wackypete2]

If you choose only the higher quality patterns as defined by autochartist you will get better results. I've used it on forex only a few years back and had some success with it. I wouldn't say it's the greatest thing since sliced bread but certainly not worthless either. If I remember correctly I only took trades on patterns where the quality, length, breakout strength, etc, were near the top of the range on the indicator they give you. For 15m patterns I only took trades in the direction of the 1H trend, for 1H patterns I only took trades in the direction of the daily trend. Experimenting with taking 1/2 off right before the target level is reached or setting stop loss to breakeven at that point might give interesting results.

Peter

 

[forker]

No, it doesn't. It's the chart patterns that are specifically identified by Autochartist.

I was referring to lamog7's statement

 

[shadowninja]

I was referring to lamog7's statement

Oh. My apologies.

 

[lamog7]

That depends on your definition of chart patterns

Def is known to everyone. Talk is cheap. Use data from 1985 - 2000 and find good chart patterns like flags, triangles, h&S, etc. Then test those pattern from 2001 - present.

 

[forker]

Def is known to everyone. Talk is cheap. Use data from 1985 - 2000 and find good chart patterns like flags, triangles, h&S, etc. Then test those pattern from 2001 - present.

Just to be clear, I am not talking of the types of patterns reflected in books and software such as this thread describes. I am referring to less obvious ones that don't come with a title.
That aside, even the ones widely known do work. How can anyone possibly conclude the validity of a price pattern without considering all those influences that dictate the market. In simple terms your tests are of singular dimension relative to an [n] dimension market. Your conclusions are therefore incorrect from the standpoint of price patterns not working. Now if you were to be more specific and say that only in terms of price, patterns have a 50% outcome. I would still argue that this is a rather incomplete view on price patterns.

 

[lamog7]

I said talk is cheap but you didn't get the message. I will say again: talk is cheap. Do you have sims that show better than 50% win rate for any patterns know to traders? if not, you can talk as much as you want and I'm afraid nobody will listen to you. You can build up your post count to 1000 fast but that doesn't prove anything.

 

[wackypete2]

I said talk is cheap but you didn't get the message. I will say again: talk is cheap. Do you have sims that show better than 50% win rate for any patterns know to traders? if not, you can talk as much as you want and I'm afraid nobody will listen to you. You can build up your post count to 1000 fast but that doesn't prove anything.

You purchased a Price Action Lab license. How are those patterns working for you?

http://www.trade2win.com/boards/trading-software/107654-price-action-lab.html#post1501280

But you're right, talk is cheap, especially yours.

Peter

 

[kalott]

backtesting?


https://temata.com/Autochartist