My dissatisfaction to the strategy tester. to the MQL developers - page 9
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Good afternoon,
and immediately a question to the developers (hopefully they will see my post):
will tick value & tick size access functions be added to synthetic formulas?
Good afternoon,
and immediately a question to the developers (hopefully they will see my post):
will tick value & tick size access functions be added to synthetic formulas?
tick value will be added to the settings
tick size can easily be derived from digits, but will also be added.
tick value will be added to the settings
tick size can easily be derived from digits, but will also be added.
Thank you for your prompt reply!
Hello, everyone. It's been a while since I've picked up a checker.
Can you advise whether the tester (or advisor/indicator in the tester) can now peep into the future? It used to be like that in the past, but what is the situation now?
This is not a complaint, nor is it a suggestion. (I mean, if the Expert Advisor shows really good results in the Strategy Tester, should we rejoice or look for mistakes?
Hello, everyone. It's been a while since I've picked up a checker.
Can you advise whether the tester (or advisor/indicator in the tester) can now peep into the future? It used to be like that in the past, but what is the situation now?
This is not a complaint, nor is it a suggestion. If the Expert Advisor shows really good results in the Strategy Tester, should we rejoice or look for mistakes?
I stopped peeping about 8 years ago when I was still in 4.
In five, use testing on real ticks, if you want maximum accuracy.
Is there any way to solve these tester problems ?
1. On stock symbols the pending cutoffs are not triggered by the last prices.
Example.
In the Depth of Market the best Bid 100 Best Ask 105.
We set Buy Limit at 101.
A new Last comes in with a price of 100. The Buy Limit order continues to hang there untouched.
2. I suspect that the tester does not consider the volume of Last deals.
I do not remember a case when the limit in the tester was partially executed.
Is there any way to solve these tester problems ?
1. On stock symbols the pending renegotiations are not triggered by the last prices.
Example.
In the Depth of Market the best Bid 100 Best Ask 105.
We set Buy Limit at 101.
A new Last comes in with a price of 100. The Buy Limit order continues to hang there untouched.
Execution of Limits at the Last price is an incomplete understanding of what is happening. In the extreme case, there may be an option to execute Limits at the Last price. But again, this is for those who have very little understanding.
Limiters in the tester do not affect the pricing. And limiters in the tester do not affect the decisions of other exchange members to send a market order. Roughly speaking, the tester is an STP market model.
Suppose your BuyLimit = 101 executed, then the flipper should not be 100, but 101. That is, the other side got SELL=101, not 100. Well and let's assume it has a fixed TP/SL. Then it will execute 1 point better/worse than it was in reality. As a result, the whole further flipper sequence has to be rebuilt. Think of it as a butterfly effect. You can't change something in the past without consequences in the future.
2. There is a suspicion that the tester on exchange instruments does not take into account the volume of Last trades.
I do not remember a case when the limit in the tester was partially fulfilled.
And that's fine here. The market side does not see your limiters in the tester. Limiters in the tester are always better than the real prices were. If that happened on the real market, the market side could have given a larger market and filled all of your limit and not partially. So again there is uncertainty - no one knows what it would be like. So again, such partial execution in the tester is only acceptable as an option, counting on the understanding of the one who includes it.
ZS The flapper bars are nonsense, dictated by historical ignorance. MT5's lack of ability to handle bid/ask bars is even more stupidity. And in general.
Forum on trading, automated trading systems and testing of trading strategies
MT4 or MT5. What are the advantages and disadvantages?
fxsaber, 2017.12.24 19:58
I think custom TFs are evil for trading. Because it's one of the arguments for maintaining life in an even greater evil - bar indicators. Terrible nastiness, let me tell you. That said, any terminal without indicators is considered dead. This is a paradox.
Execution of limiters at the price of flippers is an incomplete understanding of what is going on. In an extreme case, the option of performing limiters at the price of flippers is possible. But again, this is for those who have very little understanding.
Limiters in the tester do not affect the pricing. And limiters in the tester do not affect the decisions of other exchange members to send a market order. Roughly speaking, the tester is an STP market model.
Suppose your BuyLimit = 101 executed, then the flipper should not be 100, but 101. That is, the other side got SELL=101, not 100. Well and let's assume it has a fixed TP/SL. Then it will execute 1 point better/worse than it was in reality. As a result, the whole further flipper sequence has to be rebuilt. Think of it as a butterfly effect. You can't change something in the past without consequences in the future.
For the 10 liquid markets on the MICEX, this is a philosophical topic. It makes no difference there whether it is Last or Ask where the Buy Limit should be executed. But there are hundreds of low-liquid markets on the MICEX, where the spread may be several percent. Or even the size of a tick is several percent. There is no big sense to test illiquid instruments in the today's tester because the Limits will be executed better in the real market.
And everything is fine here. The market side does not see your limiters in the tester. Limit values in the tester are always better than they were on the real market. If this had happened on the real, the market side could have given a larger market and filled your limiters completely and not partially. So again there is uncertainty - no one knows what it would be like. So again, this kind of partial execution in the tester is only acceptable as an option, counting on the understanding of the person who includes it.
Example. Tester. You have BuyLimit at 100 in one average daily volume. Let's say you have 200 lots. In the history you had one lot at 100 pips in volume. What would you like to see in the test result? A deal of one lot or a deal of 200 lots ?
One more argument. There is no partial execution in the tester, therefore there is no possibility to test partial execution in the tester.
Example. Tester. You have a BuyLimit of 100 at one average daily volume. Let's say you have 200 lots. The history was last at 100 p in the volume of one lot. What would you like to see in the test result? A deal of one lot or a deal of 200 lots ?
I would like the mode as no partial execution in the tester and any influence of lots and flippers on the result. If other modes are optional - I would not mind, of course. But I won't use them.
I would like a mode in the form of no partial execution in the tester and any influence of lots and flippers on the result. If there will be other modes optionally - I don't mind, of course. But I won't use them.