Wish List

Quote: Originally posted by Daniel UK1

HI Peter, i know your list is full... however, i think it would be worthwhile since we now are reaching a quite high amount of indicators, to start to divide them in groups... For example momentum indicators, volume etc.. *not sure where to put the moon stuff .. but i think even a small and simple grouping would help... and not to mention the possibility to change data path for whole gsb so one can easily move from ssd to dropbox for example... Anyway, thanks for great support and development of GSB..

Quote: Originally posted by engtraderfx

Trailing Stop on close? Hi Peter, I am looking at potential longer term ideas, the gsb ATR trailing stop is intrabar exit, have you tried using closing price to trigger the exit (ie Close crosses ATR band, then close trade next open)? Sometimes filters out noise & lets trailing stop get closer to stop. Thanks Dave

Combine multiple strategies in 1 for the same symbol

Quote: Originally posted by pkotzee

I understand TS don't do well with running 2 or more strategies attached to separate charts on the same instrument, (...)

Since a few months I'm running several strategies on the same symbol, in separate charts on same TS brokerage account. The strategies have overlapping and sometimes opposing trades on most days they trade. So far, it worked fine without issues. You can use keywords like marketposition, setstoploss, setexitonclose and they work fine on the individual strategy level (not portfolio).

The main thing you need to take care of are the attached settings marked yellow and red. Yellow settings are important because your theoretical strategy position must force its live position, at least after some time the theoretical filled.

Two side issues:
1. For Eurex markets, your custom session needs to end at least a minute before exchange session for setexitonclose to work, or else your orders may be rejected.

2. If opposing trades are exited at session end, you may buy and sell at the same time, paying double slippage and commission. Such situations can be taken care of by using unmanaged orders, but it only saves you a bit of money and may not be critical depending on how often it happens. This works on the portfolio level and only one chart can do it:
OrderTicket.CancelAllOrders(GetAccountID, Symbol, false);
OrderTicket.ClosePosition(GetAccountID, Symbol);

At Interactive Brokers, issue 2. is critical because IB thinks your orders may "cross" each other and they'd be rejected.

Quote: Originally posted by pkotzee

Hi Peter, Do you think it will be possible to develop some way to combine 2, mostly uncorrelated strategies into 1 on the same symbol? Each will still have their own rules. I understand TS don't do well with running 2 or more strategies attached to separate charts on the same instrument, but since some strategies don't trade often, I think it might be worthwhile to explore this idea. Kevin Davey wrote an article about this: Article Although my easylanguage expertise is a bit limited, I'll try to do it manually as per his article and post what it looks like. Thanks

Quote: Originally posted by admin

Quote: Originally posted by pkotzee   Hi Peter, Do you think it will be possible to develop some way to combine 2, mostly uncorrelated strategies into 1 on the same symbol? Each will still have their own rules. I understand TS don't do well with running 2 or more strategies attached to separate charts on the same instrument, but since some strategies don't trade often, I think it might be worthwhile to explore this idea. Kevin Davey wrote an article about this: Article Although my easylanguage expertise is a bit limited, I'll try to do it manually as per his article and post what it looks like. Thanks Whats an issue is will you limit it to one contract, even though you have two systems. If so this is a bad idea as you get one of the big winning trades, and two of the small loosing trades. This is indirectly touched on in the newest CL video Look at low correlation of loosing trades, not low correlation of all trades

Quote: Originally posted by Carl

Hi Peter, A really nice functionality in GSB would be if we could choose a few different strategies and GSB creates a correlation matrix. Maybe a max of 10 strategies? And another one. Choosing for example two different strategies and GSB calculates the total result of the two strategies combined. Thanks.

Hi Peter,

1) Would you consider including the option for stop loss value to be a part of the search space?
2) Could you also allow "daily" bars as input. Currently we can hack it to 390 minutes, but if you start switching data1 time, it doesn't work. Daily should be daily, regardless of the intraday timing.
3) Ability to prioritize my cloud worker to my manager (on a different machine). Currently I'm trying to run jobs with my manager and I can't get my cloud worker to run my job, just other people's jobs . How is the cloud prioritization currently accomplished?

Thank you!

Quote: Originally posted by getty002

Hi Peter, 1) Would you consider including the option for stop loss value to be a part of the search space? 2) Could you also allow "daily" bars as input. Currently we can hack it to 390 minutes, but if you start switching data1 time, it doesn't work. Daily should be daily, regardless of the intraday timing. 3) Ability to prioritize my cloud worker to my manager (on a different machine). Currently I'm trying to run jobs with my manager and I can't get my cloud worker to run my job, just other people's jobs . How is the cloud prioritization currently accomplished? Thank you!

Quote: Originally posted by admin

I still feel limiting to 1 contract is not a good idea for the same reasons, but its less so if the markets are different. pa pro can cap your contracts to one, so you can test. Likely you make less $ for the same drawdown is my limited testing Best confirm what Im saying, as I could very much be wrong when its two different markets. Trading the micro contracts might be a better option. Totally understand the logic if your risk averse or low on capital, but I see it as higher risk in many ways to limit contracts numbers

Quote: Originally posted by RandyT

Regarding #3, is your cloud worker on same local lan, or remote? If remote, you'll need to exchange "share keys" between the two instances. The one running manager, and the other node running the workers. Note you also need licenses for both of these machines if that is not obvious. Share keys can be set under App settings->Workplace

Quote: Originally posted by getty002

Thanks for the fast reply Randy! I tried sharing keys and this didn't work for me, it just kept running other jobs on the worker. I tried deleting the gsb shared keys and it didn't run for me. The computers are on the same LAN - is there a different solution for this? Thank you.

Quote: Originally posted by RandyT

Quote: Originally posted by getty002   Thanks for the fast reply Randy! I tried sharing keys and this didn't work for me, it just kept running other jobs on the worker. I tried deleting the gsb shared keys and it didn't run for me. The computers are on the same LAN - is there a different solution for this? Thank you. If they are on the same lan, there is no need for share keys. And just to confirm, you have two licenses? This might be something that Peter will need to clear up for you via a Teamviewer session...

Quote: Originally posted by getty002

Quote: Originally posted by RandyT   Quote: Originally posted by getty002   Thanks for the fast reply Randy! I tried sharing keys and this didn't work for me, it just kept running other jobs on the worker. I tried deleting the gsb shared keys and it didn't run for me. The computers are on the same LAN - is there a different solution for this? Thank you. If they are on the same lan, there is no need for share keys. And just to confirm, you have two licenses? This might be something that Peter will need to clear up for you via a Teamviewer session... Yes, two licenses. Strange that the same LAN didn't work for me. However, moving my key to the top of the list did seem to do the trick. Really appreciate your help. BTW, insightful comments throughout this forum from you and others. I'm learning a lot. :-)

if you have these set these worker settings to true, or the same share key in your manager and worker - the workers should go if they are on the same lan

Quote: Originally posted by getty002

1) GSB instability is not a memory swap (RAM limitation) problem as stated in my note. It seems to occur when for example I do a create family on a worker and interact with the application directly. When I have no interaction with the session, it doesn't seem to crash. Stating that yours are stable doesn't mean that mine are. 2) My GSB is not installed to default, it's installed to Google Drive. Frankly this shouldn't matter - otherwise what's the point of specifying the directory location? 4) Having developed parallel processing applications using Open MPI on Linux, I contend that there's overhead wasted with 7 applications open. Caching in memory does not require multiple separate applications, although it was perhaps the simplest for you to implement. I believe the fact that the same data is double and triple copied on my hard drive just to manage the different applications supports my argument. 5) Where is the purchased user, non-beta release? Peter, it seems that you will continue to push forward unabated without regard to my comments on usability or stability, so I'll just be quiet now... But please keep in mind that our time has value too. And at the moment, I feel like I'm stuck in a bit of a meat grinder with many hours lost due to the above stated concerns.

Quote: Originally posted by Systemholic

hi Peter, i have a wish list. Can i suggest you consider including a new metric - Max Consecutive Loss to help user better select post build systems for further scrutiny and testing. Current metric like NP,FF,PF and NP/DD is great but they do not really show how long a system risk (risk of loss) can persist before it become profitable . I believe Max Consecutive Loss does give user a good idea and all things equal having this information can help user prioritise systems to send into favourites for further testing. Not sure where is best to place this in GSB & will leave that to your best judgement. Many thanks.

Quote: Originally posted by admin

Quote: Originally posted by Systemholic   hi Peter, i have a wish list. Can i suggest you consider including a new metric - Max Consecutive Loss to help user better select post build systems for further scrutiny and testing. Current metric like NP,FF,PF and NP/DD is great but they do not really show how long a system risk (risk of loss) can persist before it become profitable . I believe Max Consecutive Loss does give user a good idea and all things equal having this information can help user prioritise systems to send into favourites for further testing. Not sure where is best to place this in GSB & will leave that to your best judgement. Many thanks. I would like the opinions of others on this, but the request is in the job que. Hope to have it in a few weeks. Other things are being worked on now

What i do personally would want, is...

-Variance test, to be able to evaluate how the distribution looks of my strategy

-Noise Test distribution, to see that real curve is placed ok.

-Randomised OOS, to check so that OOS results are not just down to good markets and luck.

- Random systems VS real systems, to see for example top 250 systems compared to random 250 and the strategy in question..

- Random entry VS real entry, for IS and OOS, NO picture available, but same type of graph, to see real curve against 1000 random entry, i would assume one could do the same for exit.

I assume one would be able to define what is acceptable ranges for systems for these metrics, and that these would be definable in GSB as a filter perhaps..

Pictures is just to illustrate how it could look like, to give you an idea of others solution to display the metrics

GPU

Quote: Originally posted by kiwibird

Is GSB amenable to be programed to take advantage of the GPUs on Nvidia graphics cards?

Quote: Originally posted by engtraderfx

Suggestion for plotting Walkforward...when check WF result on graph when dates are limited the only way to check out of date performance seems to be to override settings but then lose the visual of before vs after, just wonder if its possible to just project the WF curve & showed as say dashed so it clear its out of sample? Would stop of a lot of back & forth? thanks Dave

Hi peter, here is example. It happens when I say build systems & WF on shorter time frame (eg say up to 2016), then want to see performance to current date I reset date & override settings again. Then need to "Use WF Parameters" to see WF on out of sample data but this removes original equity curve, sometimes changes colors too to brown. Just had an idea that could keep current current & show extended curve like so.

Quote: Originally posted by engtraderfx

Hi peter, here is example. It happens when I say build systems & WF on shorter time frame (eg say up to 2016), then want to see performance to current date I reset date & override settings again. Then need to "Use WF Parameters" to see WF on out of sample data but this removes original equity curve, sometimes changes colors too to brown. Just had an idea that could keep current current & show extended curve like so.

Quote: Originally posted by zordan

P-VALUE I would welcome a p-value test of the in-sample/out-of-sample mean returns in order to ensure OOS results are not random http://www.automated-trading-system.com/bootstrap-test/ Thanks!

Quote: Originally posted by Daniel UK1

What i do personally would want, is... -Variance test, to be able to evaluate how the distribution looks of my strategy -Noise Test distribution, to see that real curve is placed ok. -Randomised OOS, to check so that OOS results are not just down to good markets and luck. - Random systems VS real systems, to see for example top 250 systems compared to random 250 and the strategy in question.. - Random entry VS real entry, for IS and OOS, NO picture available, but same type of graph, to see real curve against 1000 random entry, i would assume one could do the same for exit. I assume one would be able to define what is acceptable ranges for systems for these metrics, and that these would be definable in GSB as a filter perhaps.. Pictures is just to illustrate how it could look like, to give you an idea of others solution to display the metrics

Hello,

The macros are fine, the methodology is fine.

The freedom to choose options is necessary and valued on my part.

Right now the freedom of choice in the Software can be improved, I do not take any credit for it, I am very grateful for this machine since it makes my work much easier and I will recommend it where the option is presented, do not hesitate.

Simple example of freedom of choice:

Quote: Originally posted by REMO755

Hello, The macros are fine, the methodology is fine. The freedom to choose options is necessary and valued on my part. Right now the freedom of choice in the Software can be improved, I do not take any credit for it, I am very grateful for this machine since it makes my work much easier and I will recommend it where the option is presented, do not hesitate. Simple example of freedom of choice:

Quote: Originally posted by Daniel UK1

What i do personally would want, is... -Variance test, to be able to evaluate how the distribution looks of my strategy -Noise Test distribution, to see that real curve is placed ok. -Randomised OOS, to check so that OOS results are not just down to good markets and luck. - Random systems VS real systems, to see for example top 250 systems compared to random 250 and the strategy in question.. - Random entry VS real entry, for IS and OOS, NO picture available, but same type of graph, to see real curve against 1000 random entry, i would assume one could do the same for exit. I assume one would be able to define what is acceptable ranges for systems for these metrics, and that these would be definable in GSB as a filter perhaps.. Pictures is just to illustrate how it could look like, to give you an idea of others solution to display the metrics

Peter any feedback in these mentioned metrics, i think they would be beneficial and appreciated.

Cheers

Quote: Originally posted by Daniel UK1

Quote: Originally posted by Daniel UK1   What i do personally would want, is... -Variance test, to be able to evaluate how the distribution looks of my strategy -Noise Test distribution, to see that real curve is placed ok. -Randomised OOS, to check so that OOS results are not just down to good markets and luck. - Random systems VS real systems, to see for example top 250 systems compared to random 250 and the strategy in question.. - Random entry VS real entry, for IS and OOS, NO picture available, but same type of graph, to see real curve against 1000 random entry, i would assume one could do the same for exit. I assume one would be able to define what is acceptable ranges for systems for these metrics, and that these would be definable in GSB as a filter perhaps.. Pictures is just to illustrate how it could look like, to give you an idea of others solution to display the metrics Peter any feedback in these mentioned metrics, i think they would be beneficial and appreciated. Cheers

Quote: Originally posted by Daniel UK1

Hi Bruce, i did not mention BA, but yes the pics to demonstrate how others are using my mentioned metrics, is from BA. I am not using BA to build any systems, just to test out the validation metrics described since i thought they had merit. And yes GSB methodology is very good, robust, works great, and is far superior to BA (not even aware they have a methodoloy), however we all have our own ways of validating and testing our final strategies, sometimes its hard to quantify whats a better or whats not metric to do this. What i do like myself and believe has most merit, is the variance test and noise test, and to quickly be able to see where my strategy is placed within the distribution, it makes sense to me personally. Is it better than anything else or what we already have, most likely not. Thanks for your input

Quote: Originally posted by admin

Quote: Originally posted by Daniel UK1   Hi Bruce, i did not mention BA, but yes the pics to demonstrate how others are using my mentioned metrics, is from BA. I am not using BA to build any systems, just to test out the validation metrics described since i thought they had merit. And yes GSB methodology is very good, robust, works great, and is far superior to BA (not even aware they have a methodoloy), however we all have our own ways of validating and testing our final strategies, sometimes its hard to quantify whats a better or whats not metric to do this. What i do like myself and believe has most merit, is the variance test and noise test, and to quickly be able to see where my strategy is placed within the distribution, it makes sense to me personally. Is it better than anything else or what we already have, most likely not. Thanks for your input we have noise test in gsb already, but I'm not convinced at all it helps us now. That and 29,30,31 min bars did help us years ago before we had all the stats and families features. Now I'm not finding it helpful and don't use it at all. You can also test on synthetic data in gsb as is. Its important to listen to ideas from users, and often we have view we hold on to - that are proven wrong. What we have is very unique and its significantly overcome the issues of having one system you trying to validate.

Quote: Originally posted by Bruce

Quote: Originally posted by admin   Quote: Originally posted by Daniel UK1   Hi Bruce, i did not mention BA, but yes the pics to demonstrate how others are using my mentioned metrics, is from BA. I am not using BA to build any systems, just to test out the validation metrics described since i thought they had merit. And yes GSB methodology is very good, robust, works great, and is far superior to BA (not even aware they have a methodoloy), however we all have our own ways of validating and testing our final strategies, sometimes its hard to quantify whats a better or whats not metric to do this. What i do like myself and believe has most merit, is the variance test and noise test, and to quickly be able to see where my strategy is placed within the distribution, it makes sense to me personally. Is it better than anything else or what we already have, most likely not. Thanks for your input we have noise test in gsb already, but I'm not convinced at all it helps us now. That and 29,30,31 min bars did help us years ago before we had all the stats and families features. Now I'm not finding it helpful and don't use it at all. You can also test on synthetic data in gsb as is. Its important to listen to ideas from users, and often we have view we hold on to - that are proven wrong. What we have is very unique and its significantly overcome the issues of having one system you trying to validate. With regards to your preference "to quickly be able to see where my strategy is placed within the distribution", I to have found that insightful in the past however when I take a good looking system and perform a WFO test the results are far from what these graphics portrayed. This could be a result of user errors on my part. One item you raised is randomizing the trade results, have you found this selection criteria really improves the system performance or robustness?

Hi Peter,

Hopefully a suggestion to consider: adding charts of drawdown, average trade and win percentage.

Update:
A system deterioration shows up in an earlier stage in the average trade chart compared to the equity chart.

Please see example charts.

Thanks.

Quote: Originally posted by Carl

Hi Peter, Hopefully a suggestion to consider: adding charts of drawdown, average trade and win percentage. Please see example charts. Thanks.

Quote: Originally posted by Daniel UK1

I think it could be intresting to have possiblity to see key metrics over time, as of now we can look at the curve to view progress over time, and we can get total metrics... but it would be good to be able to see per year, key metrics such Avr trd, %win etc... or as in Carls example, as a timeline.... in order to see if key metrics are within range... I think it would be helpful

Quote: Originally posted by Carl

Hi Peter, Thanks for adding my suggestion. But what I mean is not only the key metrics, but the average of key metrics. So for example the average trade of the 30 last trades. Please see my screenshot in my post: "ma 30 AT" The average profit factor of the last 30 trades and so on. Thanks.

Quote: Originally posted by Carl

My personal preference: Drawdown Rolling average trade of the most recent 30 trades Rolling average pearson R of the most recent 30 trades Rolling average profit factor of the most recent 30 trades Rolling average win percentage of the most recent 30 trades Thanks!

Hi Carl, yes sure, if the eld works in MC i see no reason why it would not work also for TS. PM Randy, he created the ELD, kindly shared it, and should be credited for this, and i think he runs both TS and MC so perhaps he has it already working for TS. It outputs workable fileformat for PA and MSA Portfolio same time.
Although, what MC has, i assume TS dont, is the portfolio function, meaning that i have one portfolio per parket, pre saved containing all my strategies for that market, and when i want to run my test, i just run the portfolio once.... perhaps TS has something similar you can use to avoid having to run 150 charts

The software i use, to output the final report of all strategies combined in one excel sheet, is the above mentioned MSA portfolio software.
This was the only solution i could find, would have loved to use PA that i also have, for this, but it does not have that functionality.

However i would like at some point to have PA to determine by auto, what strategies to run, but i have not had time yet to play around with this.

As a sidenote,
One thing i have been wanting to do for a very very long time, is to be able to determine from backtests, what metrics to use and how often is best to re evaluate what systems to run. The only software i found that was able to backtest historic symbol pics based on relative performance, was amibroker.
I got this project developed initially for some portfolio equity strategies i run, in order to backtest and pick symbol picks based on past relative performance. I figured i would be able to run this also for backtesting strategy evaluation, , so instead of equity "symbols" i will use my GSB strategies, and in the end i should be able to backtest different optimisation metrics for strategies and re optimisation periods.
it should work in theory.

Quote: Originally posted by Daniel UK1

NG, I shut off all my NG systems start of year, oddly enough all started to fail same time... very weird.. But it does as you say, not feel like a system fail, more a bad market condition... we see iif they pick up at some point. Last summer to winter, NG was amazing though... then it just died out.

Quote: Originally posted by Daniel UK1

NG, I shut off all my NG systems start of year, oddly enough all started to fail same time... very weird.. But it does as you say, not feel like a system fail, more a bad market condition... we see iif they pick up at some point. Last summer to winter, NG was amazing though... then it just died out.

Hey SverigeTrader , Sure that NG swing system seems interesting enough to look into some swing systems on that market from my side, good work there.
I only normally trade intraday, except equities though, but recent good work done by Peter has lured me away into some longer term system dev..

Tried intraday bonds but have not managed to get that market to work, i assume its because the little voll. Currency futures never tried.

Cheers

Quote: Originally posted by Daniel UK1

NG, I shut off all my NG systems start of year, oddly enough all started to fail same time... very weird.. But it does as you say, not feel like a system fail, more a bad market condition... we see iif they pick up at some point. Last summer to winter, NG was amazing though... then it just died out.

Hi Daniel,

My NG strats were in trouble as well.
Then I switched the entry signals. So "enter long" became "enter short" and the other way around.

Starting in August 2020, it seems NG session 0900-1430 has become more countertrend then trend following.
It looks like the trend following part is completed before the 0900 session starts, then the countertrend reaction starts in the 0900-1430 session.

Quote: Originally posted by Carl

Quote: Originally posted by Daniel UK1   NG, I shut off all my NG systems start of year, oddly enough all started to fail same time... very weird.. But it does as you say, not feel like a system fail, more a bad market condition... we see iif they pick up at some point. Last summer to winter, NG was amazing though... then it just died out. Hi Daniel, My NG strats were in trouble as well. Then I switched the entry signals. So "enter long" became "enter short" and the other way around. Starting in August 2020, it seems NG session 0900-1430 has become more countertrend then trend following. It looks like the trend following part is completed before the 0900 session starts, then the countertrend reaction starts in the 0900-1430 session.

Hi Carl, Thanks for sharing.. I would be a bit scared to flip the buy/sell like that ...
perhaps it works, but for how long ?
I am more into, discarding things when it does not work and start from scratch.. but your strategy looks great if you flip it around, so you are right on that one.

I see that you are using only 15, test 15,30 thats what worked best for me on NG.

My NG systems are still ok, but more on fading / pause mode ... until/if they start to perform again.

S though have crashed, and that market is hands off untiL new systems have been created, if possible that is.

Cheers

Quote: Originally posted by REMO755

Hello, It would be very good after having the strategy to take it to an improvement bench to be able to polish it through filters, optimization of hours, etc. It should be a module and could be called Bank of Improvements, from here those strategies that need a tweak to improve it could arise, for example trying different SF or TF, optimizing hours, eliminating days of the weeks, closing after a time interval, etc. .

Quote: Originally posted by Daniel UK1

Quote: Originally posted by REMO755   Hello, It would be very good after having the strategy to take it to an improvement bench to be able to polish it through filters, optimization of hours, etc. It should be a module and could be called Bank of Improvements, from here those strategies that need a tweak to improve it could arise, for example trying different SF or TF, optimizing hours, eliminating days of the weeks, closing after a time interval, etc. . Hey Remo, By doing that, Would you not be afraid of "over" polishing/fitting the strategy to an absolute perfection, just to just fit the historic price series better, than what is already done with GSB if not extremely careful? All the decisons before a system is picked, is, should/would/could (take your pick) be done based on results from a larger universe of systems, once a single system is picked, i would be very careful of changes to that system to improve the performance on same data that is already used. That i though, just a very personal opinion.. I do get your point of having an area to "work" on your system, after its built.. i would perhaps lean more towards validation/stress test/stats area reject or pass area ish..

Quote: Originally posted by Siem

I think it would be handy to have a "save state" option in GSB Manager. It would save the complete state of the program as a file. Part of the filename would be the version used so it could be "mycoolname.gsbstate10.0.26.69". Then if you later want to view some settings / details / graphs back of a system you made (or maybe its siblings) you can use GSBManager.62.690.exe* to open that state and GSB Manager is restored to the state it was when you used the "save state" option and do you what you want. * or whatever GSB Manager version is needed to open the file, since older versions are in the GSB folder anyway.

Quote: Originally posted by REMO755

Hello, Would it be possible to include several ranges of hours, that is, to present several options and that the algorithm can choose the best range? Example: Range 1 ---- 0900-1500 Range 2 -----1000-1500 Range 3 -----1100-1500

we have that in gsb automation recently.
a new build in in pcloud folder, but a new build coming in a few hours
This build had not been publicly released.
New spread sheet too.
has exit mode sort in it



Im hoping to have genetically chosen variation of this in GSB in next month or so

Quote: Originally posted by admin

Quote: Originally posted by REMO755   Hello, Would it be possible to include several ranges of hours, that is, to present several options and that the algorithm can choose the best range? Example: Range 1 ---- 0900-1500 Range 2 -----1000-1500 Range 3 -----1100-1500 we have that in gsb automation recently. a new build in in pcloud folder, but a new build coming in a few hours This build had not been publicly released. New spread sheet too. has exit mode sort in it Im hoping to have genetically chosen variation of this in GSB in next month or so

Edit Collection

Quote: Originally posted by REMO755

Hello, The macros are fine, the methodology is fine. The freedom to choose options is necessary and valued on my part. Right now the freedom of choice in the Software can be improved, I do not take any credit for it, I am very grateful for this machine since it makes my work much easier and I will recommend it where the option is presented, do not hesitate. Simple example of freedom of choice:

Hi,
I am also voting for something like this.
This would save a lot of time when setting indicators.

Ability to see larger chart

@ChuckNZ,
you can make it bigger by closing the other windows
ie controlR etc
see these

Forced output filter suggestion

I am sending a screenshot for a better understanding.

You can choose a departure time, now you are forced to leave at the last bar of the session.



Do we leave at the last bar of the session or do we leave at another time? Do you understand?

The best thing would be to be able to do SF with a departure time and then the algorithm would tell us the best departure time. Do we leave at the last bar of the session or do we leave at another time?

Quote: Originally posted by admin

@karish 1 good idea but i dont have resources to do this. Cant do more... 2 https://trademaid.info/gsbhelp/GSBDiagnostics.html 3 Under app settings, you can set how many cores are used, and how many wf are done at once. Too many wf at once gets painfully slow, but maxing the cores is good idea. 4 You mean font? can can adjust font scaling in windows 5 Everything should be in the docs. And there is the box down the bottom left that explains things 6 Planned and constant development done every week. No shortage of ideas, just have limited programmer resources.

1) I can open a Discord server for GSB and set you as an admin aswell as publish a link here for other users to join in, only if you agree to this of-course, i can open and sort out some channels and topics so it could be a nice and ready to use environment.

2) Thanks!, very cool, But is there any over-all knowledge-base of known indicators that wont match and you already know about that and working on it? that was what i was trying to refer to.

3) My CPU got 16 cores, i already set everything to 16 related to the cores settings, thats inside the Standalone, Worker & Manager, everyone set to 16, but still CPU/RAM wont go over 10%~..

4) The idea of how its done in EWFO would be great, until then i should just use a different screen resolution i guess..

5) Sometimes it doesn't, and the hyperlink leads to the docs where some things are not so deeply explained, or old information related to older builds.

6)

Quote: Originally posted by admin

Hi Karish I appreciate your ideas and efforts 1) if we open discord, it means there is dilution and duplicate of work on the forum. I see that as a significant negitve. Unless there was a lot of user buy in to the idea, i would not do it. 2) All indicators match on TS. Many months of work went into this, and i can now see why most other system builders will be bad at matching results. There are so many bugs / things to fix. NT has match issues, but that will be worked on more in time. But to clarify. There is an issue with Hull filter when 24 hour bars are used. (not when day session is used) and there is a error on gsb break even stop (not ts break even stop) both these bugs are on the short list to be fixed 3) you can run more WF jobs, the default is 4. You also can wf your jobs to the cloud. However if you create families from your top systems in Favorites A or b or c or d, there should not be many to wf. 4) EWFO badly needed this issue as font issues was a massive problem. Its not been significant issue on GSB 5) if there is anything you want updated, I will try to assist. Its the methodology thats the critical thing and good to follow the latest

1) I hear you, i agree.

2) Awesome!, Yeah famous builders out there advertise their software as a great advantage, last thing you know you find yourself as a non-paid beta-tester for a beautiful UI and buggy and un-matching results between the software and the trading platform in question..

3) Ill try to figure it out, thanks.

4) Thats ok, there's always a walk-around.

5) On it!

As you can see GSB is trying to remove 1 or more rules from the original strategy in order to simplify the strategy,

It will try all possible combinations first (as described above via steps),
then it will select the Best result in order to simplify the strategy.

GSB will determine if the strategy got simpler AKA better by comparing the Original stats like NP/DD etc, or what ever THE USER will decide via some settings fields,
If SimplifiedStrategy (NP/DD) > OriginalStrategy (NP/DD) then Simplify.
or any other rule or multiple rules.



we all know that the less complex the strategy, the better,
Hope it would be considered as a future feature implementation, i think it be an awesome thing to have. thanks for reading.

other rules could be added further on with further updates and or suggestions..

With this feature we can really determine if we curve fitted our strategy to the data, and just recognize where we are standing with the strategy in terms of sensitivity of parameters,

off-course we can also perform the same test with other tests like building strategies with multiple time-frames or symbols or both, this can benefit us greatly.

hope it would get considered to be added into GSB in the future, thanks for reading .

Thanks for replying back ,

I hear you, but i still think that both:

"Strategy Simplifier" & "Strategy's Optimizations Profile",

needed.

i would lean more towards "Strategy's Optimizations Profile",
Bob Pardo talks about it, here & inside the course but it is explained perfectly in a short video here:
https://www.buildingrobuststrategiesmasterclass.com/overcome...

and
System Parameter Permutation (SPP)
method was originally described by Dave Walton of StatisTrade, and is available in the paper here: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2423187

Both essentially does the same thing,
creating an optimization profile of the system in question,
running all available optimizations combinations (limited by "% Nearest" that would be set by the USER),

then if we take an automated approach,
i already explained 2 rules that would determine if the system pass/fail the test, but more rules can be added in the future,
main rules are:

"% of profitable iterations" > "90%"
thus we can be sure 90% of the iterations within our "50% Nearest" are atleast 90% profitable.

that means that we should have atleast 90% of all the iterations making 1$.

&

"Maximum degradation from the Highest iteration" < "30%"
thus we can be sure that all the iterations within our "50% Nearest" are lower than 30% from the Highest iteration.

that means that we take the highest profitable result from all of our iterations and take the lowest profitable result from all of our iterations and compare them, if we got a degradation from the highest profitable result & the lowest profitable result of higher than 30% this would be a failure.


eventually we want from this test is to see if are in the right spot with our strategy and parameters,

if we fail this test it means that our parameters are not stable from our current original position with our original strategy's parameters set.

even if we run a WF on that strategy without this test WF will show low Parameters Stability Result, and thats what we dont want..

but if we would perform this test and have passed the test,
that means our parameters are stable ("50% nearest")

then if we perform a WF with ("50% nearest") also,
we can be sure that we could have a high Parameters Stability Result.


The "Strategy's Optimizations Profile" overall is just like using the "Families" feature in GSB,
but thing is that rather waiting for 50,000 systems or more and looping this procedure and selecting the Family with the highest count of Members as it considered to be a robust set of rules with different parameters,
we could potentially if the "Strategy's Optimizations Profile" implemented do this:
Right-Click on the system in question >> Perform "Strategy's Optimizations Profile",
assuming we use "50% nearest",
and 2 rules:
"% of profitable iterations" > "90%" & "Maximum degradation from the Highest iteration" < "30%"
we can tell if the system is robust enough,
we we want to be absolutely sure, we can play with the settings or just a thought increase the "% nearest" to cover a wider range and make sure we got a robust strategy.

This feature is more direct and focused on the question.
Attached some visual images.


Thats all.
i think it would be a good addition,
thanks for reading.

@Karish
thanks for this and im looking into it.

If this feature is built, it could also be added into trademaid walkforard optimizer
I have numerous thoughts.
we dont have to do exhaustive wf. We currently have random space. But brute is fine.

Bobs example here is clear. However its problematic that he shows 2 inputs when we could have 4 to 10 inputs

does video 3 exist?

you should explore paramater analysis in GSB

@admin,

Peter related to the screenshot you posted covering 2 Input parameters on X and Y,
this was just an illustration to show a 3D view to be more presentable,

The main thing Bob talks about is what i suggested and he shows that in a 2D view,
you take all the optimizations iterations limited by % nearest of 50 for example.
form an "optimization profile",
then perform some analysis on the results of that optimization profile,

you want to see gradual results, not random profits and loses, that would mean that your parameters space of the strategy is robust and not due to a random chance, thats the most important part ("% of profitable iterations" > "90%")

as-well as the results being some what in tact with others ("Maximum degradation from the Highest iteration" < "30%")

about implementing this into EWFO, thats an idea.., but would love to see that in GSB most of all, because the transition of the strategy and its files into EWFO would cost time, it is better to just have it under our right-click inside GSB.

a right-click on a strategy and choose "Strategy's Optimizations Profile",
and GSB will perform everything behind the scenes, or maybe we could have another Tab in the center of the GUI just like we can view different charts, we could see the results of the strategy's Optimizations Profile in a Histogram form just like in the image i attached above,

i would say this is a true must have feature, it can really increase our confidence in a strategy and most likely increase the likelihood of the strategy performing well on a WF test and even the likelihood of the WF performing well into the future.


BTW:
if we could have the "Strategy Simplifier" feature as-well, we could perform the "Strategy Simplifier" test in order to see if GSB can increase the simplicity of our strategy with remove some rules that wont contribute much for the strategy,
and the "Strategy's Optimizations Profile" after that test "Strategy's Optimizations Profile" would run much faster if some unnecessary rules would be removed, thats also a creative thinking about it.


anyway, this is powerful, hope to see this in GSB in the future.. ,
it is all up to you my friend.

attached a perfect example with some fields added, something close to this example would be great to see inside GSB when implemented.

Some more small suggestions:

• For some reason not sure why, maybe it is a C# thing, but the GUI is kinda slow when there are too many rows (strategies) inside the Unique-Systems tab and click the Create Family, i also noticed that i must Remove Families before re-running the search again (play button), because the GUI just freezes and makes me lose hope every time, so i re-start GSB again with all my progress lost.., this should somehow get fixed..

• The number that shows next to the Unique-Systems (123) and Walk-Forward (123) tabs is not getting updated when deleting a strategy from the tab.

• Would be cool to have the "Entry Modes", "Weights Mode", "Entry Level Mode", "EntryLevelValue" etc as an optional list of Use/Do Not use with min/max/steps just like the "Built-In Indicators form we got already, Let the software decide what's good based on genetics etc, rather than have them fixed.

• "# of Filters" when set to 1 for example, it will be forced up on all systems.., would be cool if it wasn't, some systems will have it and some wont, again.. more freedom, let the genetics spit its results, would it be 0,1,2 etc..

• "Stops" same thing, it is currently forced on all systems..

• "Targets" also the same, it is currently forced on all systems..

• When using "Nth" it seems that the rules we set in side "Performance Filters" wont hold any merit, as if the NoTrd period of the "Nth" wont take any of its results into account.


__
That's all so far ,
thanks for reading.