High Probability Trend Strategy

Pkay · Jun 22, 2019

tommcginnis said:
If you sit down with a day's candle graph and play with the available "technical indicators" you can come up with a number of interesting 'possibles' to play with. Start with a MACD -- broadly hated, but very much NOT understood. PLAY with it.

Thanks, this is something I've been doing as well. The advised approach is to stick to the standard indicator settings and not change them but I think that this is wrong.

Do you have any tips on how to customise the MACD for a struggling trader? Thanks : )

tommcginnis · Jun 22, 2019

Pkay said:
Thanks, this is something I've been doing as well. The advised approach is to stick to the standard indicator settings and not change them but I think that this is wrong.

Do you have any tips on how to customise the MACD for a struggling trader? Thanks : )

"The advised approach" is unmitigated Idiocy. I have no idea how it came to be. There is little coverage of those who developed the indicators that are really behind it all -- Bollinger, Appel, Wilder -- NONE of them ever said anything like "ONLY USE THESE PARAMETERS." I do have a tape (somewhere) of Appel saying specifically that the numbers he was showing were specifically tailored to the market presented, and that users (the audience) should tailor their own application to their own markets/times/situations. And he says it very matter-of-factly, not in any way like he anticipated that things would've gotten frozen in the way that they have. UGH. :banghead:

"Rant over."

FWIW, two final thoughts:
• I test my lookback parameters probably every quarter.
• I have almost *never* found the default T/A settings to ever be worth a damn.

(Which means that there is frequently "fun&profit" available to time spent tweaking.

)

Pkay · Jun 22, 2019

tommcginnis said:
"The advised approach" is unmitigated Idiocy. I have no idea how it came to be. There is little coverage of those who developed the indicators that are really behind it all -- Bollinger, Appel, Wilder -- NONE of them ever said anything like "ONLY USE THESE PARAMETERS." I do have a tape (somewhere) of Appel saying specifically that the numbers he was showing were specifically tailored to the market presented, and that users (the audience) should tailor their own application to their own markets/times/situations. And he says it very matter-of-factly, not in any way like he anticipated that things would've gotten frozen in the way that they have. UGH.

"Rant over."

FWIW, two final thoughts:
• I test my lookback parameters probably every quarter.
• I have almost *never* found the default T/A settings to ever be worth a damn.

(Which means that there is frequently "fun&profit" available to time spent tweaking. )

Hey, I agree with you 110%! Even Connors, who developed the RSI, used a 2 period indicator as well as the standard 14.

Grateful if you could share one customised setting that worked for you. Maybe on the MACD : )

panzerman · Jun 22, 2019

varaamo said:
@panzerman I'd like to get your inputs. Do you think this is the right implementation of Jurik MA? if so, what you think is the right value for phase and power in the implementations? or is it something that can be discovered only through backtesting/optimization?
https://www.tradingview.com/script/nZuBWW9j-Jurik-Moving-Average/

If you insist on using moving averages, some of the newer ones like Jurik and Zero-lag by John Ehlers are a improvement over traditional ones, but only slightly.

Moving averages are just lowpass filters, and therefore the laws of physics say that the filters will have lag, otherwise known as group delay. The ideal filter has zero group delay, but that almost doesn't exist. Even half a period of group delay can seriously degrade performance of moving averages. Look into Kalman filters as well. It's a broad area of research by just may yield what you are looking for.

comagnum · Jun 22, 2019

Consistency in your equity curve pointing upward has everything to do with the trader, and little to do with the particular strategy being used.

A trader is wise to focus only on things they control, like their own consistency in following the processes of their trading plan.

All trading strategies are in sync with the market some of the time - not all of the time. The real skill lies in a traders ability to exploit their profits during times they are in sync, and minimize their loses when they are out of sync.

varaamo · Jun 22, 2019

Thank you, @panzerman . What would be the other reliable alternatives to moving averages for trend following systems?

panzerman said:
If you insist on using moving averages, some of the newer ones like Jurik and Zero-lag by John Ehlers are a improvement over traditional ones, but only slightly.

Moving averages are just lowpass filters, and therefore the laws of physics say that the filters will have lag, otherwise known as group delay. The ideal filter has zero group delay, but that almost doesn't exist. Even half a period of group delay can seriously degrade performance of moving averages. Look into Kalman filters as well. It's a broad area of research by just may yield what you are looking for.

tommcginnis · Jun 22, 2019

Pkay said:
Even Connors, who developed the RSI,

I don't know a Connors -- might be a fine guy -- but the RSI is from Welles Wilder -- it's the no-talent cousin to the much much better ADX. (Which never gets no love.)

panzerman · Jun 22, 2019

varaamo said:
Thank you, @panzerman . What would be the other reliable alternatives to moving averages for trend following systems?

Potentially any of them or none of them. It really depends on your personality, philosophy of market movement, attention span, or any other number of factors in your life.

btw, if you think prices evolve as a random walk, then all of technical analysis is charlatanism. However, even if you think it is random, Gaussian statistics do a poor job of describing market movements. I'm of the opinion that it is mostly random with brief periods of non-randomness. It's our job to sus that out, and our PnL curve is the validation of how well we do that job.

BrooksRimes · Jun 22, 2019

tommcginnis said:
I don't know a Connors -- might be a fine guy -- but the RSI is from Welles Wilder -- it's the no-talent cousin to the much much better ADX. (Which never gets no love.)

Connors is Larry Connors. He's got several books on Amazon. He likes RSI(2). https://tradingmarkets.com/

Handle123 · Jun 23, 2019

Pkay said:
Hey, I agree with you 110%! Even Connors, who developed the RSI, used a 2 period indicator as well as the standard 14.

Grateful if you could share one customised setting that worked for you. Maybe on the MACD : )

Hardly nothing works in every price action, if you have losses, price action volatility has changed. Some will call it cycling, or fib times/retracements, the moon is full, we lose for reasons most of the time of lack of knowledge or "bad luck". If new trader, you won't be accruing much "good luck" often times one trades in direction of what makes sense, but market is made up of people who trade large, not retail, and certainly do not trade as what makes sense or logical way. Having a "customized setting" would be for a time area of so much volatility and in 5 minutes this can change, if you can not read a pure chart without indicators on it, you are BAIT.

Main problems NOW than compared to the 1990's on use of indicators, there are so many quants and scientists going after price patterns, this has pulled them away from indicators, remember you want to coat tail large volume traders, they are the market. If you think your one lot going to make a difference against large trader going the other way with 500 lots during lunch, might cost you 3 ticks slippage just getting in.

It be better to study slope of price than delayed indicator that comes far short. Study price at extremes, what patterns they consistently make.