## Calculate to reflect true % return vote results

Do you like this suggestion?

# Calculate to reflect true % return Discussion

Jun 14, 2010 at 13:43
1,210 Views
16 Replies
Member Since Feb 06, 2010   24 posts
Jun 14, 2010 at 13:43
Why is trading performance calculated on a 7 day week and not on the days which are actually traded?
Member Since Sep 04, 2009   879 posts
Jun 15, 2010 at 09:43
True. ROI should be based on the trading days of the month, not on the calendar days of the month.
Member Since Oct 22, 2009   84 posts
Jun 15, 2010 at 09:58
Being sure it picks up the actual trading days of the account as opposed to just dividing it by 20 days a month.
When everything else out there is just not good enough, design your own! MIDAS
Member Since Sep 04, 2009   879 posts
Jun 15, 2010 at 10:14
excluding the Saturdays and Sundays of the month should do the job
Member Since Oct 28, 2009   1424 posts
Jun 15, 2010 at 10:25 (edited Jun 15, 2010 at 10:26)
My guess is it's tricky as different people use different brokers with different GMT offsets.
Personally think that doing some sort of scaling based on the hours might work better though.
11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.
Member Since Sep 04, 2009   879 posts
Jun 15, 2010 at 10:28
thats true Steve, but in the end we all trade only on 5 days of the week, not 7
Member Since Oct 28, 2009   1424 posts
Jun 15, 2010 at 10:36
Ah okay, I see what you're saying. I was overcomplicating my thinking. You're just on about the daily and monthly return rates in the system header which it seems simply multiply the daily rate by 30.
Personally I would assume a total of 21 days per month.
11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.
Member Since Sep 04, 2009   879 posts
Jun 15, 2010 at 10:53
by excluding the Saturdays and Sunday's of the month you get a more accurate result. i.e. June has 30 calendar days minus 8 Saturdays and Sundays = 22 trading days
Member Since Oct 22, 2009   84 posts
Jun 15, 2010 at 11:20
We are running successful strategy tests where the bot only trades 4 days a week to achieve the same results as if running 5 days, it should only judge you on what it actually does in my opinion.
When everything else out there is just not good enough, design your own! MIDAS
Member Since Oct 28, 2009   1424 posts
Jun 15, 2010 at 11:40

MIDASUser posted:
We are running successful strategy tests where the bot only trades 4 days a week to achieve the same results as if running 5 days, it should only judge you on what it actually does in my opinion.

No, it should judge you on how much your system can return in cold, hard cash in a month. That's all people care about.

What if I have an EA that only raises a trade once a month but it's 100% successful?

Are you saying that it should multiply my one trade by the number of days ( whether that is 30, 21 or 22 ) to tell me what it could make me in a month?
If you choose not to trade Friday that's your prerogative and your figures should reflect that lost opportunity.
11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.
Member Since Jul 31, 2009   1446 posts
Jun 15, 2010 at 11:58
Well, since forex is a 24 hours, 7 days a week market, that is the reason we use 30 days for a monthly calculation. There are brokers such as Oanda who do allow trading over the weekend.

As Steve pointed out, the monthly return should indeed show what return you get on average, per month. It doesn't matter which days you've traded and which you have not.
Member Since Sep 04, 2009   879 posts
Jun 15, 2010 at 14:01
Most broker servers are down for hours for maintenance work on weekends, nobody trades there. Oanda is somewhat of an exotic exception...
But I agree, whats on the table on the end of the month caunts, nothing else.😀
Member Since Nov 17, 2009   10 posts
Jun 15, 2010 at 16:20
I would also point out that there is virtually no price movement and (and very wide spreads) from New Yorks close Friday until Sydkyo open Monday.
Member Since Mar 07, 2010   257 posts
Jun 17, 2010 at 01:14

cubist posted:
I would also point out that there is virtually no price movement and (and very wide spreads) from New Yorks close Friday until Sydkyo open Monday.

oh really ? And then where do the gaps come from ?
Member Since Nov 17, 2009   10 posts
Jun 17, 2010 at 02:00
I was referring to Oanda. Gaps? I guess the lack of price movement is a result of real money players, the ones that move the market, closing shop for the weekend. At that point the the market becomes illiquid and the spreads widen in consequence.

Member Since Mar 07, 2010   257 posts
Jun 17, 2010 at 07:24
not true man forex opens in new zealand on Monday at 8 am local time (GMT+12) which is still Sunday in USA - 2-3pm EST (depends on DST)
Member Since Nov 17, 2009   10 posts
Jun 17, 2010 at 13:21 (edited Jun 17, 2010 at 13:22)
Yes, but there is no significant price movement generated on Oanda (with a few exceptions that I recall) until at least Sydney open @ 13:00 GMT (17:00 EDST)

This may explain wy NZ coming on-line Sunday afternoon has little impact:

'As FX trading has evolved, several locations have emerged as market leaders. Currently, London, England contributes the greatest share of transactions with over 32% of the total trades. Other trading centers—listed in order of volume— are New York, Tokyo, Zurich, Frankfurt, Hong Kong, Paris, and Sydney.'