Hotch's great search for the Golden Goblet!

[Hotch]

Hey all,

I'll cut to the chase because I'm shattered...

After Excel 2007 crashed on me last night, this morning i restarted programming my stratergy into excel.

Overall I probably spent 6 hours working on the project (although it could be as little as 1, I am terrible at estimating how much time i spent on things).

The stratergy was very simple, and there are many places for improvement, I really just wanted to get something to show for my time.

There are places in the actual stratergy which an be improved (I assume theoretical improvements will actually improve it atleast)
Also I didnt get around to putting in stop losses or anything like that.

The money side was just buying/selling 1 "share" of the FTSE based on my signals.

Data was from Finam Russian Brokerage Companyata Export

End of the day Id made .5%, exc commission, 43 trades (44 to close position at end of day). Max gain was 1.38%

Obviously my baby isn't anyway near close enough to being released into the big wide world, but i was pleased to see a positive result.

If results continue to go well, I might manage to get out a release in 2 weeks (8th september).

2 final notes:

1-I realise i havn't really said anything about my stratergy, that's because it's supa secret, also, I believe it is so innovative that I don't think i could explain it (not saying that it's a good stratergy!)
2-I also realise that this post is rather probably incoherentm but after 14 hours on and off in front of a screen (make that 15), and listening to "feelin' groovy" on repeat for 20 mins i am losing it.


Huzzah!

To bed!

Hotch Signing off

Love and cuddles!

xxx

 

[tomorton]

Hotch - I can't wait 2 whole weeks - get up you lazy boy and finish this strategy.

 

[yellowlion]

You ought to come to Huntington Beach. End of a long summer there are Golden Goblets aplenty that'll shred anyone's stop loss.

 

[Hotch]

Well, 2 weeks it probably will be to ge an alpha out, I'm busy this weekend, and I will also probably have to build a new computer, as I'm currently destroying my current one (top of the range a year ago). Excel is very slow (won't effect speed of order processing though).

 

[Hotch]

Playlist for today:

Mr. Wendal-Arrested Development
Fishin' 4 Religion-Arrested Development
Five a Man a Fish-Arrested Development
People Everyday-Arrested Development (the good version from the waitrose advert)
You-Bad Religion
No Woman No cry (live)-Bob Marley
Three Little Birds-Bob Marley
Buffalo Soldier-Bob Marley
Easy Skanking[*]-Bob Marley
One Love/People Get Ready-Bob Marley
I Shot the Sheriff-Bob Marley
Jamming-Bob Marley
The 59th Street Bridge Song (Feelin' Groovy)-Simon & Garfunkel
Homeward Bound-Simon & Garfunkel
Mrs. Robinson-Simon & Garfunkel
Better Open the Door-Motion City Soundtrack
Rockstar-Nickleback
Selector-Skindred
We Want-Skindred
Que Sera, Sera-Sly & the Family Stone
Teenage Dirtbag-Wheatus
Summer Brease-The Isley Brothers
Lovely Day-Bill Withers
Hey Dude- Kula Shaker
He Ain't Heavy, He's My Brother- The Hollies
Sunny Afternoon- The Kinks
Out Of The Window- Violent Femmes
A Man in A Puprle Dress-The Who
Mike Post Theme-The Who
It's Not Enough-The Who
Sound Round-The Who
We Got a Hit-The Who
Mirror Door-The Who
Play That Funky Music Whiteboy-Wild Cherry
Pappa Was A Rolling Stone-Sly & the Family Stone
Bat Out of Hell-Meatloaf
Tear Me Down-Meatloaf
Do It!-Meatloaf
Forever Young-Meatloaf
Mercury-Meatloaf

You could possibly be interested, also some crazy dutch music probably.

Go Go Procrastination :X

 

[Vrothdar]

The stratergy was very simple, and there are many places for improvement, I really just wanted to get something to show for my time.

1-I realise i havn't really said anything about my stratergy, that's because it's supa secret, also, I believe it is so innovative that I don't think i could explain it (not saying that it's a good stratergy!)

Not too fussed about knowing your strategy but even if nobody has thought of doing what you are (which is very unlikely) you should be able to explain it as you've said it's very simple.

End of the day Id made .5%, exc commission, 43 trades (44 to close position at end of day). Max gain was 1.38%

So you increased your account by .5%, before commissions over one day? Have you done any more extensive back testing? If not do you think that 2 weeks of forward testing is enough without any back testing?

 

[Hotch]

Vrothdar:

I get your point on "if it's simple it should be easy to explain", I think this is where my ineloquence with the english language becomes a problem . Guess I don't wnat to explain it either, and I know it probably has been done before, but i can't see anyone shouting about it, which suggests to me that it's successful (as every possible stratergy which fails seems to be talked about).

On Extensive testing, I will do more back testing, but there's another secret into why only 2 weeks forward testing is enough. Also, but releasing an alpha, I was planning on putting it on tradestation simulator, and so wouldn't be risking any money.

Never fear, once it all fails i will explain all, and you will mock and link me to similar strats which are inherently flawed.

UPDATE: This morning while making my indicators a bit more complex to try to improve the system, I actually made it worse over yesterday (will check against other data). I believe this is mainly due to the fact that the first stratergy relies less on stop losses etc.

Only 5 more things on the tick sheet to try out toay...

Lunch!

 

[Hotch]

UPDATE:

So I decided spread betting would be easier/more profitable.

Using £1*signal strength per point, over the last 11 months, i was down a max of £4.7 and up at the end £873, and a max up £874...so that's good right?

So Basicly

1- Have a stratergy
2- Tested it, i think the results are good.

What else do i need to do before programming it into tradestation and [\sarcasm]becoming uber rich? [\sarcasm]

Love to you all

xxx

 

[Vrothdar]

Did the period you tested it over contain various different market conditions? Have you a plan to deal with losing streaks? Have you any way of objectively determining whether your system still works as expected?

I'd suggest doing some forward testing as well to ensure that you're comfortable with it when using it in real time.

 

[Hotch]

Well, I've had a bit of a slow morning, managed to create a chart for your viewing, not sure if it shows much.

Thoughts?

IF anyone wants to hit me with some random time frames i'll do it, only reason i did years was that i had the data easy like that.

So much i want to do but excel is being rather slow. Going into the City next week so might be able to nab a bit more power for an hour or two.

P.S. Hope the Chart is attached ok. Not too sure how to do it.

 

[Vrothdar]

Chart showed up fine. My immediate thoughts are that the performance of your strategy seems to vary quite a lot from one year to the next. I'd want to take a closer look at the market conditions in each of those years to see if I can find a pattern (perhaps it only works well when the market is mostly trending).

I usually run the years end to end and look for a fairly consistent gradient for pip gains. There's some charts posted in my journal that you might find useful. Need to be careful of the scale though so that you don't lose too much detail (the charts in my journal are guilty of this). For the number of years you're using though it shouldn't be a problem.

Hopefully it's clear how this kind of visual is more useful than simply saying that your system turned £Xs into £Ys over however many years. It shows up things like that massive profit in what looks like one trade for a start.

 

[Hotch]

News!

Good and Bad

The Bad is that i am not infallible, while this was generally known, I have proof. There were a couple of mistakes in the spreadsheet (personally i blame the tragic pace, which is also my fault).

The good news is that the mistakes weren't in my favour, and so things are actually better then they appeared (woop woop!).

I have had a couple of lazy days. Havn't managed much, partly because whenever i give it a crack, excel starts being rather slow, and i start something else in the mean time and forget all about it.

I am having some trouble charting various bits, but i can present you with the actual pip chart, as the previous one is basicly wrong. While this isn't one year, it is year on year (so basicly the line is cut up and put on top of each other, you will get it if you look at it). Looks pretty consistent IMO. Doesn't look like the lines reach the same points, but they do. Also attached are some figures (yearly start, end, gain, minimum and maximum).

I presume a money chart is basicly an expodential of pips (assuming everything is reinvested), i will get onto a money chart. Also will try to get onto longest losing streak etc.

And just to note, the massive profit wasn't one trade, was just a very good streak (not that it really makes any difference).


Basicly I will try to get on with the charts for tonight. Then money management, a little more stratergy (not indicators, but how much to buy etc), and to programming!


Any thoughts what to run it on? I assumed I'd give tradestation a go. Run it on simulated for a while and then go to a prop company to stake me.

 

[Hotch]

Just a quick note.

Decided to test the stratagainst other stocks and time scales, seems to work, only interesting thing was the strat works fine for 5 min charts, and works just as good for daily charts...if you switch your buy and sell signals around.

Checked the figures, it's not a mistake in the code. Interestin eh?

 

[Hotch]

Yeah, the programming is going rather slow (as in, havn't decided what to use yet).

Did talk to Trevor Williams today though, and it was good.