IMPORTANT : BIG BUG ON SPREAD !
- This topic has 65 replies, 10 voices, and was last updated 2 years ago by
.
-
More tests needed. 🙁
So the last thing I did was using a spread of 1,9 in the Editor, with this for results on 960.000 of position : see first attachment.
Let’s say that we’re now anticipating the middle of the night where spreads can be 1.6 (twice the 0.8 I anticipate for day time). Now, should I enter 3,8 (twice the 1,9) in the editor to approach reality ?
In order to mimic that situation, I need to enter the 1,6 in the program code as well.1Spread = 1.6 // 0.8So hopefully this is not too confusing … In order to reach 0,8 I entered 1,9 in the Editor. Now I want to reach 1,9 and what to enter in the Editor for that. I start with 3,8 …
Well, see 3rd attachment. It works great.Can I now work towards a desired spread of 0,4 as well ?
1Spread = 0.4 // 1.6 // 0.8Well, not entirely because I would need to enter 0,95 (0,36 / 4) in the Editor and I can’t (that other issue). So to be on the safe side, I make 1 of that. The result is in the 3rd attachment. Thus, 0.00004 * 960.000 = 38,40 and you see that in the SpreadAmount (Orange) again.
I seem to be showing the exact same as in earlier posts, but this is not true because I off set the Spread in the Editor;
I empirically found that offset to be 1,9 over 0,8 (which seems to be 2x + 0,3 which I don’t care about) but this derived offset seems to be linear. So what I did in this post – I multiplied and divided both editor and program code numbers with the same factor, and all remains linear (and the changing of the required 0,95 for the Editor to 1 is harmless but should be rounded upwards for safety).Crazy.
And it (thus ?) does not seem to be a rounding error.Here’s another suspicious aspect :
So I require two times the expected spread plus “something”. EUR/USD is 1,13 today. 2x 0,8 = 1,6. 1,13 x 1,6 = 1,808. This is mighty close to the 1,9 needed, knowing that I can’t test for 1,808 in the Editor and it is either 1,8 or 1,9 (the 1,8 not giving the required results).
*This* now could be my own culprit, starting with taking position in my base currency of thinking, EUR, while the portfolio works in USD. So with some more “base” hoopla, this additional offset could be my own issue. Someone normally working in EUR (or in USD with USD portfolio) may not see this required additional offset. But the required times 2 should remain …?
1 user thanked author for this post.
For those who want to check for themselves … See the attachment.
Find yourself an Entry which stays flat the next bar. See the white vertical line and the one-line red to the right of the down-arrow (short in this case).
The drop in the yellow line, is the instant drop because of the denoted spread (my empirically found 1,9 for the desired 0,8 for my situation).Calculate the spread yourself from the number you entered in the Editor. Graph that (the orange one in the top in my case).
Take PositionPerf to calculate the spread as how PRT calculates it (PositonPerf shows on the 4th line here).
Multiply that with the PositionAmount. This would give you the bottom “Loss” line and it should resemble-ish the self calculated orange line.Notice that right after entry, the price remaining flat, should always provide a loss.
The Detailed Report now should show realistic results. And Yes, they will be lower than you are used to, depending on the resilience of your system/algo.
Do not forget to check (per this post) whether you are subject to the culprit in the first place. Match the top (orange) and bottom (yellow) lines;
Want to change the spread your system works with ? change it in the program for graphing (it is nowhere used for formal calculations in my case) and match the Editor number to that (observe orange vs yellow again).See it as a renewed opportunity to be able to put systems Live which not readily run into a loss because of these issues. You will do better now. 🙂
Lastly, please notice that this is a real issue only when you’re a kind of scalping. I mean, with one trade per month you won’t notice the 40 euros etc. difference on a required profit of say 1000. But if the profit is set to 50 or so to begin with (short trades), the spread is crucial.
Happy New Year !
1 user thanked author for this post.
Well, to me this is not about using integer spread numbers (pls see my posts – I never arrived at that).
But using some crazy high spread in the Editor like 1.9 in my case, also does not seems to be the solution. Not even with strong curve fitting I am able to make any profit with that anywhere.
And so there is even more going on. I mean, Live I am doing quite fine, despite the spread seems to be way off. With the spread set as I propose myself, I would be broke soon.I am a bit out of ideas as to what is actually going on.
I tend to set all back to how it was, see that the eaten spread is off by 50% or so, but make profit because my “consistent” strategy is able to cope with that. And mind you, this could be in order for anyone working so enormously long on everything. @zilliq, what do you think yourself ? others ?
Could this be stirring the pot / stoking fires without real reason ? –> Well, I know what I presented is correct. But I also know I learned to tune with that given. And personally I really can’t start all over.@GraHal, I think I feel the same as you do. It feels unsafe. Unsecure …
1 user thanked author for this post.
Find exclusive trading pro-tools on 
