Well, at least I am highly interested (thank you GraHal). But with this topic pinned to my screen, I seem to have too few time today (already yesterday). So it requires mimicking what you do, pob, and it seems that I will run into the weekend, when this can not be tested. 🙁
So at this moment I rather start spouting somewhat … and if it doesn’t make much sense, then perhaps you can still take something from it (ideas) ?
It will undoubtedly be my lack of concentration or focus, but I can’t grab the subject much. I see “entities” used, which are not conform the expressions you use. I am certainly not referring to a language problem of any sort (and otherwise always blame me) but merely to “phenomena” (entities) which do not fit. Allow me to give an hopefully justified example :
The reason why only 15 out of 40 pending stop orders on the 1-second test was actually placed, is likely to be delays when both opening and closing positions. Conditions to place an order is no longer true after 1 second, and more than 60% of the orders had at least 1 second delay.
In my book this conflicts. If you’d really still be talking about 1 second systems, then nothing can last more than one second, because PRT will have kicked you out. Explanation :
The pending orders you use will need to be repeated each 1 second, or else they are discarded. I could also say : if the broker did not respond within that 1 second, PRT will see error and throw out your system.**
but
not in Demo.
OR
Or you quite explicitly don’t talk about your Pending ENTRY orders, which …
I think are not correct because they should be Limit orders – not Stop orders.
Or you do that on purpose, because it is perfectly allowed, if you only don’t think the will be filled at the price you want (they will act as Market Orders).
… or I completely misunderstood up to don’t understand. 🙂 It could be your mechanism to test this, which I don’t grab yet (must first do it myself).
But if I am right on this one, your orders being filled will be a coincidence because you may not know in advance whether the order is at the correct side of the price (for the Stop order to match). I see you using 3 and 20 points for it and while 20 points should be OK to always be on the wrong side, 3 may not be at all (for the DAX). Mind you, the order at the right side of the price will do what you want, which is almost never.
It is also a bit in the lap of the Gods when you have Entry times at the whole minute and the half minute (or per 15 seconds which also includes the whole minute) while no single soul will ever trade on the 15 or 30 second, while many will on the whole minute. And oh, I could add to this that IG may be manipulating the prices exactly at those rarity points so you won’t notice (and just see the price changed at the whole minute). Yes, with me we can go everywhere. 😉
**): This is crucial because something which can’t happen, will now not be noticed by you and instead you could run into different issues (PRT can’t calculate the gain correctly).
On the other hand, you told that it happens in Live as well, which I don’t think you showed with screenshots and of which I may wonder how on earth you are testing *that*. Lose money ? (which exists – I do that too).
I could undermine all of my ideas above with telling that I still don’t see those delays for broker (vs PRT) response which I could easily prove by something you see below (I wrote this before I looked it up for real – haha). So this is sorted on Entries and Market orders only. All spot on the whole minute.
And to ensure I did not pick a few which coincidentally are at :00, I have a few of them (see 2nd picture) – these are just the last. Look for “Koop” (Buy) and “Bestens” (Market). Maybe I can do that for Exit as well …
… yes I can. See third picture. So that is “Verkoop” (Sell) combined with “Positie sluiten” (Exit). These are Longs and I can’t find them readily for Shorts. I do find something else though;
Picture #4 shows a mall list of erroneous trades. This is why I show the small losses as well. So, I *know* that these are failed orders, right when they started. Thus within … well, the seconds you see after the whole minute (1 or 2 or even 0 seconds). It almost looks like what you presented …
Interesting that they are all on the same 2 dates (following each other, Dec. 11 and Dec. 12).
Why this sometimes happens I could never find, but I think I have a video somewhere, testifying that because of a (too) fact price movement – while the 4 point distance which can be set by/in PRT to move with the price when a limit order is too close to the price – jumps into the too short distance after all, and the order Exits immediately. This too long sentence is to be read as : the Limit order ended up at the wrong side of the price and acts as a Market order and exits the order (because it now goes short with the same quantity while the order is long and therewith cancels the position).
Over here we say : If you can make chocolate of this, then you’re a smart guy.
(this is all too ad hoc reasoning and writing)
Sorry for my English – you must be having a hard time understanding it.
.
