I need a simple solution to a complex issue.

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Post Reply
TradingCoach
Roundtable Knight
Roundtable Knight
Posts: 176
Joined: Thu Apr 17, 2003 9:52 am
Location: Sacramento, CA
Contact:

I need a simple solution to a complex issue.

Post by TradingCoach »

I have finished with one applicaton to back test the PGO on a portfolio.
I need a simple formula to compute win loss on different contracts I have the tick value - i.e. some refer as the point value in dollars for each but the handle is different for each commodity I suppose most have 100 points in a handle so is there a way to compute some algorithm from the MINIMUM FLUCTUATION and POINT VALUE what the big handle value i.e. 1.0 move equals?
Right now I just multiply the points gained/loss with the POINT VALUE and a 100 it works for most but I am afraid the bonds and some others may yield an error. Any ideas?
Say I have 0.75 gain in crude (I know the big handle (1.0 is $1000)
the point value = 10.0
and the min fluct = .01
I just use the point value but there must be a better way to figure out the P/L for a multitude of contracts.
Forum Mgmnt
Roundtable Knight
Roundtable Knight
Posts: 1842
Joined: Tue Apr 15, 2003 11:02 am
Contact:

Post by Forum Mgmnt »

There is no easy answer to your question.

The problem relates to the numerical units as quoted by the data vendor and the contract size.

For example, some vendors quote the Japanese Yen as 85.37 while others have that same price as 0.8537. Some, Heating Oil as 85.35 (or cents), others as 0.8535 (or dollars). For Heating Oil, there are 42,000 gallons in a contract. So the Big Point Value for Heating oil in the former case is $420 (42,000 gallons / $0.01/gallon) and in the latter case $42,000 (42,000 gallons / $1.0/gallon).

Since the contract size is potentially different for each market and each data vendor quotes in different units, you will need to develop a dictionary of values for each market specific to your particular data vendor.

Some of them are not particularly easy to figure out. Try looking at the Eurodollar or Libor contract, for example.

Who are you using as your data providor? They may have this information available.

Troy at OOWDG has this information for his data. We could not get official values from CSI despite several discussions and attempts so we built our own dictionary from scratch.

- Forum Mgmnt
Bondtrader
Roundtable Knight
Roundtable Knight
Posts: 101
Joined: Mon Apr 28, 2003 11:07 am
Location: Zimbabwe

Post by Bondtrader »

Try the Australian Government Bonds (CSI symbols YTC and YTT) on the Sydney Futures Exchange. Their point value changes continuously as the underlying moves. That's right, the BigPointValue is not a constant!

http://www.sfe.com.au/content/sfe/trading/con_specs.pdf
TradingCoach
Roundtable Knight
Roundtable Knight
Posts: 176
Joined: Thu Apr 17, 2003 9:52 am
Location: Sacramento, CA
Contact:

HQuotes I find easy to use

Post by TradingCoach »

HQuotesPro and I coded to accept that data format and others that take txt or csv files.
I think I only have to account for the interest rate products as "special cases" for now..
Ted Annemann
Roundtable Knight
Roundtable Knight
Posts: 118
Joined: Tue Apr 15, 2003 7:44 pm
Location: Arizona

Post by Ted Annemann »

I found the summary page on ZapFutures's website to be extremely detailed and helpful:

http://www.zapfutures.com/html/contractspecs.shtm

They've got practically all the contracts you'd ever think of trading, and they express the "handle" / "point" / "mintick" value in a sensible way (better than the exchanges themselves).

And I don't even have an account at ZF!
Karakoram
Roundtable Knight
Roundtable Knight
Posts: 127
Joined: Wed Apr 16, 2003 11:01 pm
Location: Reno, NV
Contact:

CSI data Big point value

Post by Karakoram »

"We could not get official values from CSI despite several discussions and attempts so we built our own dictionary from scratch. "

Forum Mgmnt, I also use CSI. Would you mind e-mailing me your dictionary file, or posting it here?

Thanks!
Forum Mgmnt
Roundtable Knight
Roundtable Knight
Posts: 1842
Joined: Tue Apr 15, 2003 11:02 am
Contact:

Post by Forum Mgmnt »

Chris,

I've attached two copies of the dictionary for the liquid U.S. Futures. One for VT 1.1 and the other for VT 1.5.

The column for Big Point Value is the same and follows the currency column, which is USD for all these markets.

- Forum Mgmnt
Attachments
VT 1.1 FuturesInfo.txt
(3.68 KiB) Downloaded 999 times
VT 1.5 FuturesInfo.txt
(3.78 KiB) Downloaded 1339 times
Demon

Post by Demon »

Hi Forum Mgmnt, I have just looked through the file for V1.5 and notice that for the Dollar Index you have no correlated markets listed. With the composition of the index I would have thought it would be at least correlated with the Euro. I am very interested to know how you reached the decision to not correlate the DX with any other currency.

Best regards.

Darran :)
Forum Mgmnt
Roundtable Knight
Roundtable Knight
Posts: 1842
Joined: Tue Apr 15, 2003 11:02 am
Contact:

Post by Forum Mgmnt »

Darran,

Sorry but I didn't mention that I just added the DX and several of the stock indices and didn't update any of the correlations for those. I uploaded this to show the CSI Big Point Values not for the other values.

The DX is negatively correlated with the other currencies and the stock market indices are obviously very highly correlated with each other.

The correlations in the dictionary are only valid for those futures which made up the set that the Turtles once traded.

- Forum Mgmnt
TradingCoach
Roundtable Knight
Roundtable Knight
Posts: 176
Joined: Thu Apr 17, 2003 9:52 am
Location: Sacramento, CA
Contact:

This thread had been great help....

Post by TradingCoach »

but I am still struggling with currency P/L..
This may be a dumb quesiton but it apprears that currency P/L calc must be different from the rest due to the normal big handle division yields results in non USD where we need P/L in USD...
My progammer is in Poland and English is as alien to him as the .NET...
Kiwi
Roundtable Knight
Roundtable Knight
Posts: 513
Joined: Wed Apr 16, 2003 1:18 am
Location: Nowhere near

Post by Kiwi »

Assuming you are talking about the globex futures contracts like JY, SF, EC, BP, AD and even NE :-)

The way it is expressed is confusing but in fact each contract is in USD. So the biggies move USD12.50 per small point (for a 0.0001 point move) not some number of JY or SF or EC. So for JY the big point is 125,000 (for a 1.0000 point move)

I guess that when they say the contract size for JY is 12500000 JPY they mean that it is USD125,000.00 worth of yen when 100 yen = 1 usd. If 100 yen is worth 0.6 USD then there is still a value of 125,000 yen but the contracts value in USD is now only 75,000 USD. Despite this a 1 big point move is still going to be 125,000 USD :-)

Did that make sense? Did it even answer your question? Do you remember a crazy TV program called Soap?

John
TradingCoach
Roundtable Knight
Roundtable Knight
Posts: 176
Joined: Thu Apr 17, 2003 9:52 am
Location: Sacramento, CA
Contact:

It makes sense.

Post by TradingCoach »

Thanks Kiwi,

It does but my intended P/L now just makes a division for most everything (EXCEPT currencies) and gets the bigpoint value so the profit loss can be just computed and multiplied by the big point value. Actually (NOW) I compute the number of ticks and multiply the tickvalue(in dollars) and multiply 100
(usually there are 100 ticks in the pigpoint) except interest rate vehicles..

Now this works for Yen etc but I want to change the computation to the other way so interest rates compute right....

For YEN if divide $12.50 by 0.000001 = 12500000 !!!!! I thought that for some reason this is in Yen only to find out here that I just screwed up by the data vendor inconsistency and my own math challenge. :-)
Kiwi
Roundtable Knight
Roundtable Knight
Posts: 513
Joined: Wed Apr 16, 2003 1:18 am
Location: Nowhere near

Post by Kiwi »

You seem to be mixing the yen and usd up again. That factor of 100 (because a yen is really a cent) seems to creep in.

Divide USD12.50 by 0.0001 (the tick that Globex and CSI is using for yen) to get USD125,000 per contract big point. I know that with IB they use 0.000001 for their tick but I dont understand why they do that - it just makes it harder for your IB Front Ends and Graphing Packages.

John

Acutally just looked at the globex quotes and they skip the decimal point all together and quote it as 9997. lol, no wonder we get confused.
Bondtrader
Roundtable Knight
Roundtable Knight
Posts: 101
Joined: Mon Apr 28, 2003 11:07 am
Location: Zimbabwe

Post by Bondtrader »

Maybe the thing to do is - - - - - - - - hire a coach :lol: to teach you about it.
TradingCoach
Roundtable Knight
Roundtable Knight
Posts: 176
Joined: Thu Apr 17, 2003 9:52 am
Location: Sacramento, CA
Contact:

I did

Post by TradingCoach »

My programmer in Poland suppose to know this, but you get what you pay for. :cry:
Post Reply