Disliked{quote} How can you fall asleep with discovery of this magnitude? {image}Ignored
Diversified Trend Trading Approach
- #3,682
- May 19, 2018 4:00pm May 19, 2018 4:00pm
- Joined Oct 2009 | Status: Exploiting psychology of the crowds | 1,426 Posts
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
- #3,683
- May 19, 2018 4:04pm May 19, 2018 4:04pm
- | Commercial User | Joined Apr 2013 | 4,366 Posts
Ernest Chan in his book Algorithmic Trading: Winning Strategies and Their Rationale (Wiley Trading) postulates that some markets are more mean reverting and some are more divergent. For example if you would create spread between Brent and WTI, this would be convergent or mean reverting market. We know that stuff happens that such market goes haywire and deletes money on your account. As it was with LTCM. That's why we stick with divergent and positive skewed strategies. But if some markets show more convergent nature like EUR GBP, AUD NZD, AUD CAD,...
It might be best to wait until the EA is fully developed and then we can attempt to slay the beast with full riguour and our mighty swords :-(
Cheers C
- #3,688
- May 19, 2018 6:39pm May 19, 2018 6:39pm
- | Commercial User | Joined Apr 2013 | 4,366 Posts
Note to the Interested
ATR is a bad estimate to use for position sizing as Pip said. The EDTT used the SDC for determining volatility adjusted position sizing (not ATR) and we need to incorporate a far more effective method of adjusting for volatility in this KAMA approach. This will be an extension to this testing.
The KAMA has proved very effective in squeezing the juice out of trending periods ....provided you can keep away from the normal daily churn. My gut tells me that Pip was already fully aware of this.....and hence down the track we need to start looking at more efficient measures than that deployed in this preliminary version......such as Theil-Sen etc. which was suggested. Pip's home grown trend following indicator probably has significant legs and under a portfolio would probably fly like the clappers....so treat this for a while as a work in progress.....but we are not here to be spoon fed an indicator.....we are here to create our own and in the process actually understanding them before we apply them.
The first step is however to demonstrate that the choice of an indicator such as KAMA is effective because it attempts to discriminate between noise and signal. This is not a price following indicator but a signal assessment indicator......and there is a big difference in their effectiveness as a trading tool
So for those that think ALL indicators are lagging and bad.....I beg to disagree.....think again. Once again you need to test all assumptions you might hear on this forum.
Not all tools are alike. Pick your tools carefully after you understand them.
ATR is a bad estimate to use for position sizing as Pip said. The EDTT used the SDC for determining volatility adjusted position sizing (not ATR) and we need to incorporate a far more effective method of adjusting for volatility in this KAMA approach. This will be an extension to this testing.
The KAMA has proved very effective in squeezing the juice out of trending periods ....provided you can keep away from the normal daily churn. My gut tells me that Pip was already fully aware of this.....and hence down the track we need to start looking at more efficient measures than that deployed in this preliminary version......such as Theil-Sen etc. which was suggested. Pip's home grown trend following indicator probably has significant legs and under a portfolio would probably fly like the clappers....so treat this for a while as a work in progress.....but we are not here to be spoon fed an indicator.....we are here to create our own and in the process actually understanding them before we apply them.
The first step is however to demonstrate that the choice of an indicator such as KAMA is effective because it attempts to discriminate between noise and signal. This is not a price following indicator but a signal assessment indicator......and there is a big difference in their effectiveness as a trading tool
So for those that think ALL indicators are lagging and bad.....I beg to disagree.....think again. Once again you need to test all assumptions you might hear on this forum.
Not all tools are alike. Pick your tools carefully after you understand them.
- #3,690
- May 20, 2018 4:22am May 20, 2018 4:22am
- | Commercial User | Joined Apr 2013 | 4,366 Posts
DJI30 Medium Range Test Complete - 1 Jan 2016 to 20 April 2018
EDTT surrenders to the power of KAMA
As expected KAMA nailed it and exceeded my expectations....but despite the glorious equity curve there were some hair raising losses along the way.
Now onto the tougher stuff to see how KAMA takes the rocky road. Does EDTT have any dry powder left to fight KAMA?
C
EDTT surrenders to the power of KAMA
As expected KAMA nailed it and exceeded my expectations....but despite the glorious equity curve there were some hair raising losses along the way.
Now onto the tougher stuff to see how KAMA takes the rocky road. Does EDTT have any dry powder left to fight KAMA?
C
- #3,691
- May 20, 2018 5:00am May 20, 2018 5:00am
- Joined Oct 2009 | Status: Exploiting psychology of the crowds | 1,426 Posts
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
- #3,694
- May 20, 2018 6:27am May 20, 2018 6:27am
- | Commercial User | Joined Apr 2013 | 4,366 Posts
{quote} I listened to Mr. Alvarez's interview and I have such notes: 1. These stop loss findings are very dependent on given strategy - he repeated it many times. In different strategy other stop losses would be better or worse than in this one. 2. He researched stocks from Russell 2000: stock charts are long-biased and especially growth stocks: The Russell 2000 Index is a small-cap stock market index of the bottom 2,000 stocks in the Russell 3000 Index. 3. Test everything: when we will have the tools, we need to test different stop loss options...
- #3,696
- May 20, 2018 6:29am May 20, 2018 6:29am
- Joined Jan 2016 | Status: I am not a trader | 1,238 Posts
- #3,697
- May 20, 2018 6:41am May 20, 2018 6:41am
- | Commercial User | Joined Apr 2013 | 4,366 Posts
Copernicus, Ramadas and others, this music masterpiece is for you ... BN, thread lurker https://www.youtube.com/watch?v=2u-P...&start_radio=1
- #3,698
- May 20, 2018 11:13am May 20, 2018 11:13am
- Joined Oct 2009 | Status: Exploiting psychology of the crowds | 1,426 Posts
Copernicus, Ramadas and others, this music masterpiece is for you ... BN, thread lurker https://www.youtube.com/watch?v=2u-P...&start_radio=1
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
- #3,699
- Edited 2:39pm May 20, 2018 2:04pm | Edited 2:39pm
- Joined Oct 2009 | Status: Exploiting psychology of the crowds | 1,426 Posts
I had to complete 1h/4h manual backtest on the better trending market: USD TRY.
Using Karma 1h with 4h filter.
Risk 1% per position, compounding.
We will another option also in higher time frames.
You know, if we will have capacity problems with smaller timeframes. It is nice problem to have, for sure.
Used yet old technique (2 days old - time files fast in these parts) of one position.
2 years
MAR is 34.09/7.74=4.40
Using Karma 1h with 4h filter.
Risk 1% per position, compounding.
We will another option also in higher time frames.
You know, if we will have capacity problems with smaller timeframes. It is nice problem to have, for sure.
Used yet old technique (2 days old - time files fast in these parts) of one position.
2 years
MAR is 34.09/7.74=4.40
Cut short your losses. Let your profits run on. David Ricardo (1772-1823)
- #3,700
- Edited 9:39pm May 20, 2018 8:34pm | Edited 9:39pm
- | Commercial User | Joined Apr 2013 | 4,366 Posts
I had to complete 1h/4h manual backtest on the better trending market: USD TRY. Using Karma 1h with 4h filter. Risk 1% per position, compounding. We will another option also in higher time frames. You know, if we will have capacity problems with smaller timeframes. It is nice problem to have, for sure. Used yet old technique (2 days old - time files fast in these parts) of one position. 2 years MAR is 34.09/7.74=4.40 {image} {image} {image}
About the Donchian Break - I have my reservations still about trading on the Donchian break as it is a key point of market structure which is typically used as a method of market/support and resistance by many traders and investors......hence algos. We need to offset our trade entry in the direction of the breakout to reduce the number of whipsaws as this is a key potential reversal point for mean reversion....like we do with EDTT. Furthermore when the Donchian 2000 starts to deteriorate and moves to a new major market structure point, it loses it's logical integrity. A better method would be to use say a month lookback for a higher high or a lower low (which is equivalent to a 2000 lookback on M30) and then offset from this defined zone.
Just looking at ways to perhaps further compensate for the Sumo's....like our requirement in EDTT for the short term momentum to exceed the longer term momentum on breakout. We might also need these little tweaks in momentum surge on entry to assist in reducing the Sumo's.
Looking forward to the EA mate as this crunching is painful manually. We will be able to sweep through this process far faster with an EA....but what I really want to see is how the portfolio curve stacks up with the process and compare that against EDTT.
We also need to keep in the back of our minds that the last few years has been a miserable trending regime in general. Imagine if volatility returns with a vengeance. That's when we upgrade to Chiron's :-)
Cheers C