Enormous Points

[FetteredChinos]

Now, the question is....

is this a curve fit?

been toying with seasonalities in the FX market lately, and noticed something quite remarkable...

firstly, the stats....(1991-2003)

Pips: 128,087 (this is not a typo)
Trades: 2033
Wins: 1208 (59%)

Pips per trade: 63


Now, the entry rules..

Longs:- on the 3rd month of each futures quarter (ie march, june, sept and dec) enter a long on the close of each trading day.

Shorts:- on the 1st month of each futures quarter (ie Jan, april, july, october) enter a short on the close of each trading day.

Exits:- Exit all trades at the end of the month they were entered...

NO STOPS...


equity never dipped more than approx 20,000 pips over this period. as a result, you could probably trade this at £1 a point per £40,000 of equity...

at that conservative stake size if returns approx 20% ROI per annum...

obviously entries etc can be improved, but this was just something i noticed and thought i would share with you lot.

FC

 

[JonnyT]

I wish you hadn't!!!

20,000 pips drawndown. As McEnroe used to say 'You cannot be serious'

This method also ties in with the Lunar charts...

Need I say more?

 

[FetteredChinos]

yes, but you have to look at drawdown in relation to overall gain...

if you wanted to cut the drawdown, then just trade on the opening day of each month...

that more amenable for you,sir?

 

[trendie]

FC,

being as the trade lasts approx 1 month, could this be traded more efficiently using options ?
( known downside, but possibly higher percentage increase to upside, bearing in mind time-value decay )

I am no expert ( just see my journal ), but using options may allay the downside scaredy-cats. ( just joking )

do you have any stats on which times in the month there is minimal volatility, and which times seem to have greatest volatility ?

 

[JonnyT]

Sorry but 20,000 pips drawdown and only 20% per annum

JonnyT

 

[trendie]

FC,

there is a sort of pattern emerging from your studies.
namely;
a: not to have stops, or only have big, drop--dead stops.
b: exit after specific time-elapsed date, or at certain target, rather than use some indicator.

just an observation, but you seem to be moving towards a Jake Bernstein-type seasonal probability-trades type scenario.

although you are much more entertaining. ( and cheaper )

keep up the good work.

 

[the blades]

Now, the question is....

is this a curve fit?

FC

well, I can see why you cut the data off in 2003!

but seriously, keep up the good work, it's great food for thought for many others, myself included.

UTB

 

[pkfryer]

I've been doing lots of system design and testing last few months on my newly finished analyser which took a while to finish. One thing I found is if you dont use stops you can pretty much make incredible wins consistently if there is even a slightly upward drift to the market.

I have hit fools gold with a few systems, got consistent 60+% wins with 3 win/loss ratio with one system that I called the Ema squeeze. Using two emas, when they converged you entered and diverged you exited. No stops either and really crappy rules really.. but none the less it worked very well because one of the rules only happens during an up trend and one happens only during a down trend. It theoretically made a lot but the drawdowns were totally ****.

Also, another thing I found was a bug where i set the stop order accidentally .5% above the entry level. So it was a fixed target above the range, without realising it. I was experimenting with an incrementing stop level that narrowed the gap between the stop order over time regardless of price action. But it was bugged :| . So it was a small amount above the trade with no stop below prices. 95% of the time with most stock the small target got hit eventually. So, 95% winners and theoretically it would have made billions. Only problem is sometimes it would be in a gigantic drawdown for several years, but the market eventually recovered on most occasions.

I really should get around to coding in Margin calls!

The sort of systems that dont have stop losses and dont care about draw downs are as useless as used toilet paper. Looks alright in theory but you couldn't ever trade them in real life. I've got 6 systems that give theoretically incredible results, but with no stops and unlimited potential drawdown are actually worthless for trading.

 

[rog1111]

Personally the way I try to compare these sorts of mechanical systems is to take the ratio of the average profit and loss per annum /max drawdown over the entire period tested.

If your ratio is much less than 3 then it will be a VERY hard system to trade as your drawdowns will probably kill you, both mentally and financially, even if you are trading very small £/$ per point. A ratio > 5 should be fine, 8 is fantastic. There is also a strong case for taking the more pessimistic ratio of your worst 12 months rolling profit over the period and dividing this by the max drawdown over the whole period.

So in FC's example, the system makes 128087 pips in approx 146 months with a max DD of 21020 pips.

So, ratio 1 is 128087/146x12 / max DD = 10528 / 21020 = 0.5

.....which IMO would be untradable, for me at least.

One only has to look at the period from 17/07/92 - cummulative pips = 23245, to
16/09/02 - cummualtive pips = 2225, to realise how hard it would have been to survive this, 9 months of trading wiped out in a couple of months, and if this had come near the start.......

Thanks though for posting FC, perhaps there are some tweaks that can be made ?

rog1111