I've been reading the thread "MC 8.5 walk forward optimizer is FLAWED" (https://www.bigmiketrading.com/multicha ... lawed.html) from BigMikeTrading. The thread was created by treydog999. Thanks a lot to him.
I only use sometimes the WFO functions in Multicharts. But now I read, the results can not be trusted 100%. So is it a waste of time using walk forward optimization?
@Multicharts team:
1. What is the current status? Are any improvements in MC 8.7 realized?
2. What bugs still exits in MC 8.7 regarding WFO?
3. What workaround does exist to overcome the problems?
4. Are the WFO issues adressed for the next version of MC?
WFO is one of the highlights MC offers, so it is a pity that the problems still exist for so many versions. Thanks again to treydog999 for informing at BigMikeTrading.
MC 8.5 walk forward optimizer
- JoshM
- Posts: 2195
- Joined: 20 May 2011
- Location: The Netherlands
- Has thanked: 1544 times
- Been thanked: 1565 times
- Contact:
Re: "MC 8.5 walk forward optimizer is FLAWED"
The WFO can certainly be improved, so I'm looking forward to hearing the answers.
(Thanks for creating this thread btw, swz168).
(Thanks for creating this thread btw, swz168).
-
- Posts: 323
- Joined: 07 Jan 2007
- Location: Lahore, Pakistan
- Has thanked: 64 times
- Been thanked: 57 times
Re: "MC 8.5 walk forward optimizer is FLAWED"
Robert Pardo, on page 299 of the second edition in his excellent book, The Evaluation and Optimization of Trading Strategies, writes:
By the way, there is a $20 walk-forward analyzer software which summarizes both the profit, or loss, in the walk-forward, and the Walk-Forward Efficiency ratio for a user of any ability.
Khalid
It may be a reflection on my abilities, rather than any shortcoming(s) , of MultiCharts walk forward optimizer that I have found it rather tedious to the point of ‘better not bother with it’ to calculate either the walk-forward profit, or the Efficiency, when using MultiCharts.How does the Walk-Forward Analysis uncover optimization excesses and statistical failure? The most important way is to determine whether it made a profit in the walk-forward. This is its major contribution to the art of trading strategy development. The second important measure is at what relative pace or degree of efficiency did it make a walk-forward profit. This is the role of Walk-Forward Efficiency.
By the way, there is a $20 walk-forward analyzer software which summarizes both the profit, or loss, in the walk-forward, and the Walk-Forward Efficiency ratio for a user of any ability.
Khalid
- JoshM
- Posts: 2195
- Joined: 20 May 2011
- Location: The Netherlands
- Has thanked: 1544 times
- Been thanked: 1565 times
- Contact:
Re: "MC 8.5 walk forward optimizer is FLAWED"
These are still good questions two weeks later. Did you get a response yet Swz168?(...)
I only use sometimes the WFO functions in Multicharts. But now I read, the results can not be trusted 100%. 0. So is it a waste of time using walk forward optimization?
@Multicharts team:
1. What is the current status? Are any improvements in MC 8.7 realized?
2. What bugs still exits in MC 8.7 regarding WFO?
3. What workaround does exist to overcome the problems?
4. Are the WFO issues adressed for the next version of MC?
- swz168
- Posts: 120
- Joined: 30 Sep 2010
- Location: Germany
- Has thanked: 47 times
- Been thanked: 32 times
Re: "MC 8.5 walk forward optimizer is FLAWED"
Unfortunately no.Did you get a response yet Swz168?
At least for MC 9.0 Beta 1 some other improvements is targeted according to PM: https://www.multicharts.com/pm/viewissu ... no=MC-1087
But I think those improvements are not related to the problems reported by treydog999.
Hopefully, we still will get an answer for the questions above.
- JoshM
- Posts: 2195
- Joined: 20 May 2011
- Location: The Netherlands
- Has thanked: 1544 times
- Been thanked: 1565 times
- Contact:
Re: "MC 8.5 walk forward optimizer is FLAWED"
I agree, but that PM request (as I understand it) is somewhat vague about the actual, concrete implementation(s). For example, adding a fitness function from TS (as mentioned in that PM request) is redundant since we can code this ourselves with custom criteria. So we'll have to see what actually gets implemented.At least for MC 9.0 Beta 1 some other improvements is targeted according to PM: https://www.multicharts.com/pm/viewissu ... no=MC-1087
But I think those improvements are not related to the problems reported by treydog999.
There are, by the way, some good suggestions for the WFO made in this thread. Perhaps that the programmer who's going to implement that mentioned PM request for MC 9.0 can also address the issue raised in this thread and implement something from that other thread 'in one go'.
---
Edit: Somewhat off-topic but still relevant in this context: now that NT8 still hasn't been released yet (and with no release date communicated), this is an excellent time for MultiCharts to pick up some features for algorithmic trading (WFO, bar magnifier in portfolio). And with IQBroker coming this quarter (which has great backtesting: e.g., already multi-currency support, portfolio trading and a 'bar magnifier' for portfolio backtesting) it's really the time to devote more programming hours to the needs of algorithmic traders.
- JoshM
- Posts: 2195
- Joined: 20 May 2011
- Location: The Netherlands
- Has thanked: 1544 times
- Been thanked: 1565 times
- Contact:
Re: "MC 8.5 walk forward optimizer is FLAWED"
Wow, still no response from MultiCharts Support. What's up with that?
In my view, this thread is about MultiCharts users looking for tips on how to best work with the WFO so that problems are avoided, and a request for an update about future features.
I'm not sure why that doesn't deserve a response. It surely cannot be considered offensive or highly critical.
In my view, this thread is about MultiCharts users looking for tips on how to best work with the WFO so that problems are avoided, and a request for an update about future features.
I'm not sure why that doesn't deserve a response. It surely cannot be considered offensive or highly critical.
@Multicharts team:
1. What is the current status? Are any improvements in MC 8.7 realized?
2. What bugs still exits in MC 8.7 regarding WFO?
3. What workaround does exist to overcome the problems?
4. Are the WFO issues adressed for the next version of MC?
- Andrew MultiCharts
- Posts: 1587
- Joined: 11 Oct 2011
- Has thanked: 931 times
- Been thanked: 559 times
Re: "MC 8.5 walk forward optimizer is FLAWED"
Hello Everyone,@Multicharts team:
1. What is the current status? Are any improvements in MC 8.7 realized?
2. What bugs still exits in MC 8.7 regarding WFO?
3. What workaround does exist to overcome the problems?
4. Are the WFO issues adressed for the next version of MC?
WFO is one of the highlights MC offers, so it is a pity that the problems still exist for so many versions. Thanks again to treydog999 for informing at BigMikeTrading.
I personally worked with treydog999 on his WFO case, so let me answer each question:
- At the moment, no improvements are scheduled for any particular MC version.
- In the latest versions of MC (8.7 and 8.8 ) there are no issues with WFO. I also confirm that i have proved couple of times to treydog999 that the issues he reported are not problems with MC. WFO works correctly and in some cases it is easy to prove,while in some cases it is hard to prove. However, at the moment there is no known issue with WFO in MultiCharts software.
- If you experience any issue, please report it to MultiCharts Support directly so it can be addressed in a timely manner.
- As I said previously there is no issue with MC, so nothing is addressed for next versions of MC.
As soon as WFO is to be improved in any particular version of MC, the feature request will be assigned for it.