DislikedThanks for the thread and ea. Any recommendation for an mt5 boker for the demo. This wil be my first mt5 venture. So I need easy to start unrestricted demo. ThanksIgnored
Do you load this file into the experts folder?
- #127
- Aug 28, 2023 11:53am Aug 28, 2023 11:53am
Hello all. My "DOSI" take "DailyRebound Portfolio Trader" is now available in MQL5 MartetPlace. Trades and allows testing complete portfolio from single EA instance.
https://www.mql5.com/en/market/product/104032
Been trading variant of DOS for couple years and this is where I have ended to. Besides convenient portfolio trader my main change to original strategy is to enter couple of hours (configurable per pair) before the market close. I have found out that this is when the pullback/exhaustion/reversal starts in most cases. (As bonus it eliminates issue with high spreads). Due to that I dont make decision based on actual daily bar size but based on "virtual daily bar" which is the move during 24 hours before the trading time.
If you want to test my version it should be run on H1 timeframe and make sure all traded pairs are in your MarketWatch. Leave comments to MarketPlace if you happen to test it. Demo version is available for free to test in the StrategyTester.
https://www.mql5.com/en/market/product/104032
Been trading variant of DOS for couple years and this is where I have ended to. Besides convenient portfolio trader my main change to original strategy is to enter couple of hours (configurable per pair) before the market close. I have found out that this is when the pullback/exhaustion/reversal starts in most cases. (As bonus it eliminates issue with high spreads). Due to that I dont make decision based on actual daily bar size but based on "virtual daily bar" which is the move during 24 hours before the trading time.
If you want to test my version it should be run on H1 timeframe and make sure all traded pairs are in your MarketWatch. Leave comments to MarketPlace if you happen to test it. Demo version is available for free to test in the StrategyTester.
- #131
- Edited 11:31am Sep 23, 2023 11:16am | Edited 11:31am
@Forex37
I recommend 5000EUR or USD. With smaller account size risk starts to increase. With 2.5K risk effectively almost doubles. Its still doable but expect almost double DD compared to 5K. You can try to simulate with free demo version from the MarketPlace how it would behave with smaller balances. You need to likely increase Max Account Loss parameter to allow more room to breath
@yoriz
Thanks for testing my EA quite thoroughly and discovering issue with 1min OHLC bar testing in the Strategy Tester in some setups! There is now new version in the MarketPlace which can be more reliably tested in the tester and also runs with less CPU load on live environments
@all
There is 3 things I am still going to look as further improvements
1. Try volume based filter to try to avoid big DDs. I have noticed that large daily moves do not necessarily rebound at the end of the day if relative volume is still much higher in that evening than it typically is in the same evening hours. I will try to build filter testing that finding/hypothesis. Not sure if this will improve things further...
2. Exit at strength also when in loss. When X trades deep in the basket exit after strong pullback also at loss. If there really started strong trend against the basket this would exit at the first pullback of that trend. Will reduce profits as loss baskets then happen every now and them but avoids catastrophic losses
3. Add multiplier for lot size if swap is positive for the direction and symbol. Its "almost" nice to be multiple trades deep in basket if swap is generating money every night
So some baskets could turn out to be "carry" trades. This cannot be backtested as there is no historical data for swaps. Tester always uses current swap. But I basically want this feature for my own use and for forward testing.
I recommend 5000EUR or USD. With smaller account size risk starts to increase. With 2.5K risk effectively almost doubles. Its still doable but expect almost double DD compared to 5K. You can try to simulate with free demo version from the MarketPlace how it would behave with smaller balances. You need to likely increase Max Account Loss parameter to allow more room to breath
@yoriz
Thanks for testing my EA quite thoroughly and discovering issue with 1min OHLC bar testing in the Strategy Tester in some setups! There is now new version in the MarketPlace which can be more reliably tested in the tester and also runs with less CPU load on live environments
@all
There is 3 things I am still going to look as further improvements
1. Try volume based filter to try to avoid big DDs. I have noticed that large daily moves do not necessarily rebound at the end of the day if relative volume is still much higher in that evening than it typically is in the same evening hours. I will try to build filter testing that finding/hypothesis. Not sure if this will improve things further...
2. Exit at strength also when in loss. When X trades deep in the basket exit after strong pullback also at loss. If there really started strong trend against the basket this would exit at the first pullback of that trend. Will reduce profits as loss baskets then happen every now and them but avoids catastrophic losses
3. Add multiplier for lot size if swap is positive for the direction and symbol. Its "almost" nice to be multiple trades deep in basket if swap is generating money every night
So some baskets could turn out to be "carry" trades. This cannot be backtested as there is no historical data for swaps. Tester always uses current swap. But I basically want this feature for my own use and for forward testing.
- #133
- Dec 10, 2023 12:46pm Dec 10, 2023 12:46pm
Looks like some issue with broker symbol settings. Perhaps you need to use symbols with some prefix or postfix ? Seems that broker does to allow opening new trades on those symbols ("EURUSD") but only closing them. I am not actively checking and reacting to those errors. Double check that symbols are right for your broker or is pre or postfix need to be added to all of them ?
I think very many people actually have issues to properly backtest my multi-symbol EA. I realized that MT5 is actually not able to load automatically historical data for all the secondary pairs traded. MT5 automatically downloads missing data only for the host/chart symbol symbol: 
I have always tested all pairs individually before running full portfolio and hence not seen that issue. That guaranteed that there is needed historical data for all symbols you need to first run EA one pair at a time selecting matching Symbol in tester settings (above) and in EA settings (below)
While tedious benefit is that then you see results of each pair then individually as well. After running each individually you are able to successfully test full portfolio as well as data will be available. I will have mode in next version to do this data download for backtesting easier.
I think very many people actually have issues to properly backtest my multi-symbol EA. I realized that MT5 is actually not able to load automatically historical data for all the secondary pairs traded. MT5 automatically downloads missing data only for the host/chart symbol symbol:
I have always tested all pairs individually before running full portfolio and hence not seen that issue. That guaranteed that there is needed historical data for all symbols you need to first run EA one pair at a time selecting matching Symbol in tester settings (above) and in EA settings (below)
While tedious benefit is that then you see results of each pair then individually as well. After running each individually you are able to successfully test full portfolio as well as data will be available. I will have mode in next version to do this data download for backtesting easier.
- #134
- Edited 1:24pm Dec 10, 2023 12:57pm | Edited 1:24pm
RVOL filter introduction will actually greatly improve this strategy and my EA ! It seems like too good to be true actually 
Above trades now 20 pairs and would actually backtest wihout large DD from 2013- even though above test is from 2016-
I use RVOL to enter the trades in the evening only if relative volume is low enough. And adjust lot size based on the RVOL as if volume drops low pullback is very likely. I found clear relation in trade success % and RVOL value. If RVOL stays high all evening trade better be skipped as large move is likely caused by large news which elevates volumes for a longer time and move might continue.
I will publish 2.0 version with RVOL soon to the MQL5. I need to test / figure out right default Stop Loss levels for it as I dont want to publish settings without any Stop Loss as past performance does not guarantee future and I dont want people to potentially lose lots of money with my default settings....
I had these plans above:
1. Try volume based filter to try to avoid big DDs. I have noticed that large daily moves do not necessarily rebound at the end of the day if relative volume is still much higher in that evening than it typically is in the same evening hours. I will try to build filter testing that finding/hypothesis. Not sure if this will improve things further...
This ended up being large improvement and will be in the next version
2. Exit at strength also when in loss. When X trades deep in the basket exit after strong pullback also at loss. If there really started strong trend against the basket this would exit at the first pullback of that trend. Will reduce profits as loss baskets then happen every now and them but avoids catastrophic losses
Could not get this working well without significantly reducing performance. Best alternative was to exit when 2 period daily RSI was high enough (>70-80) and we were at loss. Then baskets were closed at first pullback when in some cases trend continued against us after the pullback. But it also often hurt. So this is still further investigation if there is way to do this without destroying the performance.
3. Add multiplier for lot size if swap is positive for the direction and symbol. Its "almost" nice to be multiple trades deep in basket if swap is generating money every night So some baskets could turn out to be "carry" trades. This cannot be backtested as there is no historical data for swaps. Tester always uses current swap. But I basically want this feature for my own use and for forward testing.
Left out for now due to lack of time and the fact that this cannot be backtested
Above trades now 20 pairs and would actually backtest wihout large DD from 2013- even though above test is from 2016-
I use RVOL to enter the trades in the evening only if relative volume is low enough. And adjust lot size based on the RVOL as if volume drops low pullback is very likely. I found clear relation in trade success % and RVOL value. If RVOL stays high all evening trade better be skipped as large move is likely caused by large news which elevates volumes for a longer time and move might continue.
I will publish 2.0 version with RVOL soon to the MQL5. I need to test / figure out right default Stop Loss levels for it as I dont want to publish settings without any Stop Loss as past performance does not guarantee future and I dont want people to potentially lose lots of money with my default settings....
I had these plans above:
1. Try volume based filter to try to avoid big DDs. I have noticed that large daily moves do not necessarily rebound at the end of the day if relative volume is still much higher in that evening than it typically is in the same evening hours. I will try to build filter testing that finding/hypothesis. Not sure if this will improve things further...
This ended up being large improvement and will be in the next version
2. Exit at strength also when in loss. When X trades deep in the basket exit after strong pullback also at loss. If there really started strong trend against the basket this would exit at the first pullback of that trend. Will reduce profits as loss baskets then happen every now and them but avoids catastrophic losses
Could not get this working well without significantly reducing performance. Best alternative was to exit when 2 period daily RSI was high enough (>70-80) and we were at loss. Then baskets were closed at first pullback when in some cases trend continued against us after the pullback. But it also often hurt. So this is still further investigation if there is way to do this without destroying the performance.
3. Add multiplier for lot size if swap is positive for the direction and symbol. Its "almost" nice to be multiple trades deep in basket if swap is generating money every night
So some baskets could turn out to be "carry" trades. This cannot be backtested as there is no historical data for swaps. Tester always uses current swap. But I basically want this feature for my own use and for forward testing.Left out for now due to lack of time and the fact that this cannot be backtested
This way to trade still works...
After a longish absence from FF I decided to visit here as I finally published new version of my EA to MQL5 MarketPlace.
It has now RVOL filter and and position size adjustment to avoid trades or take smaller trades when relative volume is high (this indicates there is still pressure to move price to the direction it has already travelled a lot). RVOL is measuring if volume of evening hour is higher or lower than at the same time on the few previous days.
Also added mechanism to skip DCA trades on the large baskets if the trend is strong. Those trades are then entered when the trend clearly slows down. This is combined with mechanism to exit at loss if basket is large, delayed averaging trades are done and there is short term pullback in the trend. This is almost like automatic stop loss which exits in pullback of the large trend. Sometimes it would have been reversal allowing to get out in profit but based on testing it is beneficial to get out in assumed pullback and sometimes the the loss.
EA is at least initially cheap as chips still as I check if there is any interest and if I should spend any effort publishing it or EAs in general. And it is free to try in the strategy tester. I have been trading this way now from 2020 and I have put quite a bit of effort to finetune and test the original approach.
Backtest from 2013- to today (Risk Multiplier = 1.0, default)
And last couple months forward testing the latest version (Risk Multiplier = 1.5). There is this DD period where high volatility killed lots of grid martingale EAs in the latter part of the July. There was some heat also for this but not too bad. Some of it was NZDJPY which was the only JPY pair I had in. I now removed that too as JPY pairs are really best left for breakout and trend trading nowadays. 
Search for "DailyRebound" from MQL5 MarketPlace if you want to try. Dont want to put link here... Video is old but EA is updated finally.
After a longish absence from FF I decided to visit here as I finally published new version of my EA to MQL5 MarketPlace.
It has now RVOL filter and and position size adjustment to avoid trades or take smaller trades when relative volume is high (this indicates there is still pressure to move price to the direction it has already travelled a lot). RVOL is measuring if volume of evening hour is higher or lower than at the same time on the few previous days.
Also added mechanism to skip DCA trades on the large baskets if the trend is strong. Those trades are then entered when the trend clearly slows down. This is combined with mechanism to exit at loss if basket is large, delayed averaging trades are done and there is short term pullback in the trend. This is almost like automatic stop loss which exits in pullback of the large trend. Sometimes it would have been reversal allowing to get out in profit but based on testing it is beneficial to get out in assumed pullback and sometimes the the loss.
EA is at least initially cheap as chips still as I check if there is any interest and if I should spend any effort publishing it or EAs in general. And it is free to try in the strategy tester. I have been trading this way now from 2020 and I have put quite a bit of effort to finetune and test the original approach.
Backtest from 2013- to today (Risk Multiplier = 1.0, default)
And last couple months forward testing the latest version (Risk Multiplier = 1.5). There is this DD period where high volatility killed lots of grid martingale EAs in the latter part of the July. There was some heat also for this but not too bad. Some of it was NZDJPY which was the only JPY pair I had in. I now removed that too as JPY pairs are really best left for breakout and trend trading nowadays.
Search for "DailyRebound" from MQL5 MarketPlace if you want to try. Dont want to put link here... Video is old but EA is updated finally.