My Study in Optimization

Gianluca · Sat, 29 Dec 2018 10:42:33 +0000

Hi guys, another year has passed, i want to start a discussion to the developement of automatic tradings and optimization. Most of us discovered in many ways (sometimes with our blood $ :p ) the danger of overfitting of the optimization of strategies. I’ve startet my journey of automatic trading with a discover of a strategy on this forum, the strategy was made by a talented user named Wyngs, i’ve found the strategy here: https://www.prorealcode.com/prorealtime-trading-strategies/wings-resistance-breacher-dax-30m/ At the beginning i was shocked and excited about it and i started that in real account. The first months i earned 2k euros or more, but since feb the strategy started to loose money and i stopped it. I had an idea i tried to check on a 200k bars graph and the differences from the original post where big, so i discovered that the strategy was overfitted only for 100k bars, and i had an idea. I tried to overfit that only for the last 6 months, then for 100k bars and then for 200k bars. Here is the result on my demo account 1 year later. Many times most of people that i met in my first year of trading told me that a good strategy to perform well need to be backtested and work well for a long periods. But with this little and inaccurate study i want to start a discussion that i hope can help many of us to find a good way to manage the automatic algos. In this little confrontation of those strategies my conclusion are that the atomatic algos doesn’t need a too long and overfitted optimizazion but they nee a constant an monitorated update every 6 months. Specially in theese years of change on the financial market, and specially for the 2019 that could be a another year of changes. Here the photos of my study. ps. I’m sorry for my englis but as you can see i’m just Italian :p

This topic has 22 replies, 5 voices, and was last updated 5 years ago by CKW.

Viewing 8 posts - 16 through 23 (of 23 total)

← 1 2

01/10/2019 at 4:18 AM #88282

CKW

GraHal wrote:

CKW wrote:

If I understanding the most recent variable from WFA report, it does not include the market behaviour between 30/7/18 – 9/1/19.

30/7/18 – 9/1/19 is an OOS period, so variables set / optimised in the IS period (27/11/17 to 30/7/18) are applied (without any further optimisation) to prices in the OOS period.

Hope that helps?

Agreed but my doubt is the variable to be used on live (i know this has been asked for many times). Let me illustrate it for better understanding.

I read the WFA related Guideline / Post from this forum that we should use the most recent variable from WFA analysis (which is variable C). But this is contradict with my understanding of WFA result-> “In order to have a similar gain as shown in historical OOS (s), I should find the best performance variable from last 8 months of data (from today) and use it on the next 5 months. After 5 months, i must repeat this process.”

If i use variable C to live, this meant i am expecting variable C will work for 10 months (OOS(3) + Current Live). This is the main doubt i have when i started to apply WFA for my systems. By the way (from my observation), i don’t expect a good system to have major change from variable C & variable D. If it does, it’s risky for me.

i think this is somehow related to conclusion from Gianluca: “In this little confrontation of those strategies my conclusion are that the atomatic algos doesn’t need a too long and overfitted optimizazion but they nee a constant an monitorated update every 6 months”.

May i have your opinion, shall we use variable C for live or variable D? (if variable C & why?)

br,

CKW
Participant

Topics: 22
Replies: 87
Been thanked: 8 times

01/11/2019 at 5:55 PM #88412

GraHal

CKW wrote:

shall we use variable C for live or variable D?

I think Variable value D is more logical?

But isn’t your 5 month OOS arbitary anyway … simply because you chose 3 Repetitons of the WFA?

Should there not be more iterations of the WFA (with, for example 5 WFA Repetitons) in order to determine the most logical period to reoptimise? (I’m not saying I do all this … just what we should do! 🙂)

GraHal
Participant

Topics: 204
Replies: 8428
Been thanked: 2046 times

01/12/2019 at 4:58 AM #88418

CKW

GraHal wrote:

I think Variable value D is more logical?

Yes. It’s more logical. But from Nicholas guideline, he recommended variable C if I am not wrong. Hope to get his view

GraHal wrote:

But isn’t your 5 month OOS arbitary anyway … simply because you chose 3 Repetitons of the WFA?

GraHal wrote:

Should there not be more iterations of the WFA (with, for example 5 WFA Repetitons) in order to determine the most logical period to reoptimise? (I’m not saying I do all this … just what we should do! )

For sure we should always pick best WFA result among different level of repetitions (5 which is the max for my platform), I got 3 repetitions w best perf for above example. This example was simulated based on 30m time frame, WFA 70%/30%, 100k bars so 5 months OOS.

br,

CKW
Participant

Topics: 22
Replies: 87
Been thanked: 8 times

01/12/2019 at 11:02 AM #88422

GraHal

CKW wrote:

Nicholas guideline, he recommended variable C if I am not wrong

I would only go with Variable Value C (most recent IS period) if the period looked typical as far as price movement / fluctuations.

For example if there was a massive spike or fall then I would not use Variable Value C on Forward Test as the optimiser will have chose best value for Variable C to give max profit for that (not typical) most recent IS period and so likely will not give the best performance going forward.

Also to then further optimise over the period including the most recent OOS period and so end up with variable values for (your) period D has to better than period C alone (more data from very recent) … provided of course that the variable values don’t differ wildly from period C WFA optimisation values.

In my opinion … there will always be a element of subjective judgement when deciding on which IS period WFA optmisation values to go forward with and Nicolas recommendations are for guidance and not rigid adherence?

Please feel free to offer alternate views anybody?

GraHal
Participant

Topics: 204
Replies: 8428
Been thanked: 2046 times

01/12/2019 at 12:19 PM #88430

CKW

GraHal wrote:

I would only go with Variable Value C (most recent IS period) if the period looked typical as far as price movement / fluctuations. For example if there was a massive spike or fall then I would not use Variable Value C on Forward Test as the optimiser will have chose best value for Variable C to give max profit for that (not typical) most recent IS period and so likely will not give the best performance going forward. Also to then further optimise over the period including the most recent OOS period and so end up with variable values for (your) period D has to better than period C alone (more data from very recent) … provided of course that the variable values don’t differ wildly from period C WFA optimisation values.

Agreed. It’s Case by case. .

GraHal wrote:

In my opinion … there will always be a element of subjective judgement when deciding on which IS period WFA optmisation values to go forward with and Nicolas recommendations are for guidance and not rigid adherence?

no, no, Nicholas recommendation is perfect. This discussion is just counter check if anyone doing the same or advice not do it

CKW
Participant

Topics: 22
Replies: 87
Been thanked: 8 times

01/12/2019 at 12:28 PM #88431

GraHal

CKW wrote:

no, no, Nicholas recommendation is perfect.

Wow not even a smiley there, @Nicolas has to see this! 🙂

GraHal
Participant

Topics: 204
Replies: 8428
Been thanked: 2046 times

01/14/2019 at 10:01 AM #88522

Nicolas

Use variable D for the next live trading period. Through different topics and explanations about WFA across the website, I might have made a confusing English explanation somewhere, let me know where and I’ll fix it 😉

Nicolas
Keymaster

Topics: 81
Replies: 20828
Been thanked: 4742 times

01/17/2019 at 12:19 PM #88909

CKW

It’s clear now. Thanks Nicholas !

CKW
Participant

Topics: 22
Replies: 87
Been thanked: 8 times
Author

Posts

Viewing 8 posts - 16 through 23 (of 23 total)

← 1 2

Find exclusive trading pro-tools on