Discussion of article "Extract profit down to the last pip" - page 23

[[Deleted]]

fxsaber:

I waited. The trade is halted.

Isn't that premature?

[fxsaber]

trader_number_one:
Isn't it premature?

I don't know.

[fxsaber]

The search toolkit from the article has improved

Running Optimisation.

At the end of the best pass data are taken and from them are formed (set ranges of input parameters) several tasks for optimisation.

All optimisations from item 3 are carried out.

The best passes are taken from all optimisations from item 4 and saved as sets for portfolio trading.

It turned out to be a very powerful market scanner and TS adjuster. TC sources are not required for such manipulations.

and became a TS adjuster as well.

[traveller00]

Good afternoon. Since the article is focused on forex, I did a similar thing with searching for a suitable source of tick data for the stock exchange. I ran through 4 main sources that I managed to find

h real account of Otkritie broker.

The test was run on all overlapping instruments that are provided. For 2 weeks from 1 November 2019 to 15 November 2019.

On futures from worst to best erinrv-finam-Opening-zerich. On a fairly small number (~5-10 out of 70) sometimes Finam overtakes Otkrytie, but not by much. And only once Otkritie was better than zerich and that within the margin of error and on an insignificant instrument, where the potential profit was less than 1000 pips. In other words, zerich is in the lead here, although the gap is not dramatically strong.

As for shares, Finam falls out, as the data there are only futures data. The remaining 3 are arranged from the worst to the best erinrv-zerich-open. Sometimes erinrv is ahead of zerich, but not much. But the gap of Discovery is very noticeable: by an order of magnitude, or even by two. Here, Otkrytie is unequivocally in the lead, the gap is unambiguous, concrete and strong.

As for currency instruments (at least for that very small part that is represented there), Finam falls out again, because the data there is only futures data. Of the remaining 3, zerich and erinrv are on the same level, somewhere one instrument is better, somewhere the other. And Otkritie is getting ahead again. It is noticeable, but not by orders of magnitude as in the previous case. Here, Otkritie is unequivocally in the lead.

I hope the research will be useful to someone. If anyone knows other sources of tick data, please write, I will also add them to the test. I am interested in exchanges, not forex, with ask and bid prices. Thank you.

[fxsaber]

traveller00:

I hope the research will be useful to someone. If anyone knows other sources of tick data, please write, I will add them to the test as well. I am interested in stock exchanges, not forex, with ask and bid prices. Thank you.

Thank you for sharing the information. If there is no error in the study, then it appears that the data in different sources do not match. It looks strange, as the exchange is a centralised trading platform, i.e. everything should coincide, if the collectors do not skip and filter anything.

If I understood correctly, you were running a script of potential profit (mentioned in the article) with zero commission. If the commission was set different for each of the sources, then the mismatch is immediately explainable.

I am very glad that the approach from the article has aroused constructive interest in such a study. I wouldn't even look at exchanges like this myself, as I was sure that everything was the same for everyone (except for technical problems).

ZЫ QScalp History Data-format I don't know. If you share the parser, it would be great.

[Dmitriy Skub]

traveller00:

Good afternoon. Since the article is focused on forex, I did a similar thing with searching for a suitable source of tick data for the stock exchange. I ran through 4 main sources that I managed to find

h real account of Otkritie broker.

The test was run on all overlapping instruments that are provided. For 2 weeks from 1 November 2019 to 15 November 2019.

On futures from worst to best erinrv-finam-Opening-zerich. On a fairly small number (~5-10 out of 70) sometimes Finam overtakes Otkrytie, but not by much. And only once Otkritie was better than zerich and that within the margin of error and on an insignificant instrument, where the potential profit was less than 1000 pips. In other words, zerich is in the lead here, although the gap is not dramatically strong.

As for shares, Finam falls out, as the data there are only futures data. The remaining 3 are arranged from the worst to the best erinrv-zerich-open. Sometimes erinrv is ahead of zerich, but not much. But the gap of Discovery is very noticeable: by an order of magnitude, or even by two. Here, Opening is unequivocally in the lead, the gap is unambiguous, concrete and strong.

As for currency instruments (at least for that very small part that is represented there), Finam falls out again, because the data there is only futures data. Of the remaining 3, zerich and erinrv are on the same level, somewhere one instrument is better, somewhere the other. And Otkritie is getting ahead again. It is noticeable, but not by orders of magnitude as in the previous case. Here, Otkritie is the leader unconditionally.

I hope the research will be useful for someone. If anyone knows other sources of tick data, please write, I will also add them to the test. I am interested in exchanges, not forex, with ask and bid prices. Thank you.

Strange results. Tick streams should coincide absolutely. Both in time and value. Since the source is the same.

I once compared Open with Finam - complete coincidence (it was at the time of checking).

[traveller00]

I agree that the results are non-standard. That's why I decided to share it, maybe it will be useful for someone, to save a week of time for modifying scripts and software to suit themselves.

Links to tick histories I took mainly from here real account as standard, it picked up the history itself. For the other 3 sources I took the programme from here https://github.com/StockSharp/Qsh2Bin/releases  It is an open-source (mostly), but, as it turned out, it is not without its faults. So I had to tweak and rebuild it. What had to be corrected, I wrote here https://stocksharp.com/articles/322/konvertatsiya-istoricheskih-failov-qscalp-v-format-stocksharp?tid=322& page=3 And a little tweaked it for myself, so that it picked up parameters from the command line and parsed without any windows. Accordingly, I modified the script for myself so that after downloading from online it runs this converter and then picks up the received csv. It's a crutch, of course, ideally I should have parsed the qsh binary from the script at once, because the format description is available at https://www.qscalp.ru/store/qsh.pdf and there's nothing complicated there. But it turned out to be easier and faster, I haven't managed to rewrite it properly yet. If you need, the programme and edited sources can be posted, but I will tell you at once, they are all in crutches, written for myself just to convert and that's all. The speed is also not great, but for a few hours last year all (about 70) characters parses.

[fxsaber]

traveller00:

I agree that the results are out of the ordinary.

If you take out some part of the full tick history, the potential profit will drop.

I think MT5 tickhistory was compared with the one that the exchange officially publishes. There was a complete similarity.

So everything speaks in favour of the fact that QScalp does not log everything, hence the discrepancies.

ZЫ This information should probably be passed on to the users of this product.

[traveller00]

Perhaps it's QScalp. If anyone has other sources of tick data or has an opportunity to run it on a real account of another broker, it would be interesting to compare.

[[Deleted]]

Trade resumed?