Big Figure Days

[TWI]

Interested to know if anybody accounts for figure releases in their mechanical systems.
For example.
A system like JonnyT's Spot On would probably take a short this morning on the downside break in EURUSD.
Is this a good idea ahead of non-farms or is it prudent to wait until the release before doing anything?
Naturally if the figure is the "right" way around it would be advantageous but then this "if" possibly makes things more of a gamble than a trade?? Maybe over time it evens out each way..I don't know, has anybody done this sort of figure analysis?
Personally I have not factored figures into any mechanical system but I certainly take great account of it with the scalping, ensuring I do not have any potentially dodgy orders before or through any major announcement.

 

[rog1111]

FWIW I have looked at the non-farm payroll days results for the past 13 months (not a huge sample I know) for the JonnyT type break out system & GBP, 70 pip stop. At first glance results don't seem to be anything out of the ordinary ie -38, -22, NT, +56, -70, +54, +70, +103, -70, NT, -70, NT, NT and today NT. Bearing in mind that this is with a 2 hour time restriction on entry, so often the markets may be waiting for the news release and nothing much happens early on. An all day entry study would be more interesting (will post later), as would FOMC meetings etc.

rog1111

Interested to know if anybody accounts for figure releases in their mechanical systems.
For example.
A system like JonnyT's Spot On would probably take a short this morning on the downside break in EURUSD.
Is this a good idea ahead of non-farms or is it prudent to wait until the release before doing anything?
Naturally if the figure is the "right" way around it would be advantageous but then this "if" possibly makes things more of a gamble than a trade?? Maybe over time it evens out each way..I don't know, has anybody done this sort of figure analysis?
Personally I have not factored figures into any mechanical system but I certainly take great account of it with the scalping, ensuring I do not have any potentially dodgy orders before or through any major announcement.

 

[jimvt]

I've been looking at this sort of thing recently. I have a mechanical support/resistance system for cable and I was curious to see if it would improve if I kept out of the market around announcements.
So far I have just programmed in (in Tradestation) the 1:30 pm announcement dates so that if there is an announcement that day I won't enter a position between 1:00 and 2:00.
I haven't so far filtered the announcements as to whether they are (considered) important or not - I just included all of them.
The result is (obviously) a reduction in the total # of trades but a modest increase in the average trade size and PF. My conclusion is that the improvement is signifiant (i.e worth doing) if not dramatic.

This is for a short term supp/res scalp strategy though. I haven't tried it on a breakout system that might have a longer holding period.

 

[TWI]

Thanks for the feedback.
Rog - those results are interesting as if you stayed out on non-farm days you would have ended up about square (+13) so maybe it does just cancel out over time. think it would be interesting to see on non farm days if there is a reversal or continuation of morning price action. That will provide the answer for any breakout system regardless of timeframe. Something like (Close- 13:29hrs price) + (13:29hrs price - Open) .
Jim - do you just enter each date seperately into EL code in TS or is there a way to reference some sort of calendar?

 

[jimvt]

TW

There may well be a better way but I typed them in manually.

I have emailed the part of the code to you.

 

[rog1111]

Further to my earlier post, a summary of non farm payroll days for GBP 1-8 BO system with no 2 hour entry restriction, 70 pip stop, is as follows :

July 03 -38
Aug -22
Sep -70
Oct +56
Nov -70
Dec +54
Jan 04 +70
Feb +103
Mar -70
Apr +178
May -70
Jun -70
Jul -70
Aug (today ) -70

Seems like lots of losers. I'll check on a later time range (eg 01h30 to 13h30) for these days and report back, along with the same for EUR.

rog1111

 

[rog1111]

OK, I have re-run the test, but this time I thought it would be more interesting by allowing trade re-entry if stopped out at -70 pips. So if we were long and stopped out, then we go short at the point of exit, with no slippage allowed for. Although under "normal" conditions results deteriorate with re-entries, on non farm days, it gets interesting. This is only a small sample and other big news days would be interesting to look at too. Here are the results : -

July 03 -38
Aug -22
Sep -70 / + 8
Oct +56
Nov -70 / + 19
Dec +54
Jan 04 +70
Feb +103
Mar -70 / + 203
Apr +178
May -70 / + 9
Jun -70 / - 70 / - 48
Jul -70 / + 71
Aug -70 / +122

rog1111

 

[rog1111]

Here are the results for EUR with a 60 pip stop, with re-entry allowed :

July '03 -33
Aug -60/0
Sep -60/+125
Oct +99
Nov +100
Dec -60/+35
Jan '04 +13
Feb -60/+113
Mar -60/+135
Apr +179
May -60/-60/-60/+140
Jun -60/-60/+55
Jul -60/+111
Aug -60/+168

Hmm...

rog1111

 

[shogun]

rog1111 you are of course assuming that you get filled at your stop , which is very difficult if the mkt gaps over your order , i am currently looking into forex options as a possible hedge around non farm payrolls,

costs seem to be about the same between igindex and refcofx of about 30p each side , so if you were long euro you would buy a at the money put for 30p before the payrolls , if the mkt rallied 80 pips your put would be worthless but your long would be up 80 pips, if the mkt fell 80 pips your put would be in the money 100 pips ( 80 + time value ) and your open position would be down 80 and if the mkt stayed flat you could sell your put for about 15-20 p ,

has any one else looked into this type of hedging around figures,

fwiw i trade a end of day system with my signals at 2100 hrs and find the non farms & greenspan a pain in the **** good when your on the right side but bad if you are wrong, but i find over time they tend to even out,

 

[drb2]

I would never personally put take a position before a big number coz if the number somes out differently to what u expect you wont be able to close position and you'll end up way offside

 

[TheBramble]

I would never personally put take a position before a big number coz if the number somes out differently to what u expect you wont be able to close position and you'll end up way offside

Unless you take both sides with a reasonable stop?

 

[rog1111]

shogun

Yes this is true, but the purpose of the exercise was simply to try to decide whether these days have looked especially unusual or bad in the past via backtesting. Conclusion - with wide stops, no.

rog1111

rog1111 you are of course assuming that you get filled at your stop , which is very difficult if the mkt gaps over your order ,

...............

fwiw i trade a end of day system with my signals at 2100 hrs and find the non farms & greenspan a pain in the **** good when your on the right side but bad if you are wrong, but i find over time they tend to even out,