Xmas Project

[Tuffty]

"is the recent increase in volatility of 1997 onwards here to stay? 7 years is a sustained rise, or is it going to tail off and return to longer term levels??"

Wasn't SETS (electronic order book) introduced in the UK Equity market about then?. Market structural change. Just a thought.

 

[FetteredChinos]

thanks tufty. thats kinda what i assumed. it seems that breakout strategies worked long term up until about 1997 too (the ones that i have tested) and then fell over. btw,do you know when some of the other Europeans went electronic? Kraftwerk excluded.

only one way to find out if things work, and thats to test em.

i might do a bit more testing over the coming weeks to try to find simple improvements, and then go live perhaps in the new year. might even crank up the journal again...

cheers fellas.

 

[Tuffty]

FC, Can't remember. I think the UK was slow of the mark as the London Stock Exch was worried about competition from Continental Europe where I think some exchanges were already electronic. (Such a long time ago, I was working for a data vendor at the time so I should know!!)

 

[FetteredChinos]

ok thanks Tufts.

much appreciated. i'll go digging for info. not that it makes too much difference in hindsight. might explain a few things though..

 

[TWI]

According to the study (covering S&P 500 from 1976 to 1985) if the week starts XXX then there is 62% probability of a reverse on Thursday (XXXO) and a 38% probability of continuation (XXXX). Then for Friday, if XXXO then a 52% probability of reversal (XXXOX) and 48% continuation (XXXOO) OR if XXXX then 62% for continuation (XXXXX) and 38% reversal (XXXXO).

I tested this and sorry to say it no longer makes sense. Indeed, if you had traded it since 1996 you would now be 200 full points per lot in the money but this only maches the peak of 200pts profit in Oct 2000

 

[FetteredChinos]

ah knackers. thanks though TW. didnt think it worked substantially well.

if you were to hold for more than the next day though, it does have a degree of merit (see my Andy Kaufmann thread for more details.)

fc

 

[TWI]

Running a quick curvefitting optimisation I notice following
Best results achieved if following
Fri/Mon/Tue up days sell Wed
Fri/Mon/Tue/Wed up days, buy Thursday
Hold each for 7 days.

It is still sh1te however.

 

[FetteredChinos]

that i do not disagree with.

nasty drawdowns etc...