Fully automated futures trading

globalarbtrader said:
I'll revisit this subject in my next monthly blog post, in November.

GAT
More...

Would you mind giving a brief hint in the meantime?

I tried many things. Priors as average correlation, priors as corr of 0.5, none seem to show any order, still shrinkage factor is all over the place.
 
test_user said:
Would you mind giving a brief hint in the meantime?

I tried many things. Priors as average correlation, priors as corr of 0.5, none seem to show any order, still shrinkage factor is all over the place.
More...

I wouldn't waste your time trying. I doubt I will be able to recover the original numbers and I will be using a better method in the new post (based on parametric distribution of correlation estimates). And none of this will make a blind bit of difference to how much money you or I make.

GAT
 
globalarbtrader said:
1. Did you frame your earlier optimization problem using training set, validation set and testing set?

No. I don't use a 'validation set' per se.

2. Let's say if we have 10 years of data, what would be the best split for training set, validation set and testing set?

I always use expanding out of sample windows in annual chunks. So for the first year, we can do no testing. Then we fit a model based on year 1 data, and test it throughout year 2. Then we fit a model based on years 1 and 2 data, and test it on year 3. And so on. That means the split is always N year training set (where N is as large as possible without cheating) and 1 year test set.

We can tweak this with rolling windows, or some kind of exponential weighting where more recent data gets weighted more highly.

'3. For my explanation below, let's assume first 8 years of data to training set, 1 year of data to be validation set and last year of data to be testing set
4. Training set: Bootstrap (independent or block) for first 8 years of data (assume training set), with each bootstrap sample 10% of the training set. Find the optimal set of rule weights in each sample to max sharpe?
5. Validation set: Should the average optimal weights (across all samples) derived from training set be considered vis-a-vis the validation set e.g. Sharpe ratio in training set shouldn't differ from validation set by 5%. If it differs, find a set of weights with reasonably high sharpe ratio (not necessarily the highest) but doesn't differ from validation set by 5%'

'e.g. Sharpe ratio in training set shouldn't differ from validation set by 5%.': Hahaha, good luck with that! The inherent randomness of SR means that over 1 year even a perfectly good rule could easily end up with a completely different SR. You'd need to eithier make the SR bands very wide (i.e. over 1 year the 95% uncertainty range of a 0.5SR trading strategy is -1.46 to 2.46) or make the validation set decades long (which means you're wasting a huge amount of training data that isn't affecting the model, the more valuable data as well which is more recent).

'6. Testing set: Not used in model fitting but used in backtested equity curve?

Yes, I think so.

'Steps 3 to 6 repeated with expanding window.'

'My concern would be the validation set. How long should it be and should validation be set up like a slightly modified version of K-fold cross validation in machine learning problems? E.g. using K-1 folds of data but excluding data before validation set to prevent lookahead bias.'

I think the whole validation set idea here is flawed, because there is too much noise in the data. You are better off using the entire set of data that is in the past at a historical point, and doing a fitting process on that which is robust i.e. which accounted for the number of years of data and the noise.

GAT
More...

Hi @globalarbtrader

Thanks for your comprehensive reply and noted your objections in using validation dataset for fitting a trend following model. And I guess it differs from conventional machine learning problems which are more data-driven (requires validation set for the 'feedback') than idea-driven.

Very useful pointers that could potentially save me a lot of time and 'steer' me towards the right path :)

Jirong
 
Hi everybody,
sorry for a very basic futures trading mechanics question, but I just need to confirm it for myself:
Is buying a non-FX futures contract denominated in a foreign currency also a bet on that currency or not (not considering margin money)?

e.g. how should I calculate my P&L if my base currency is USD and I bought 1 3ktb contract a month ago, and since then usd.rkw spot exchange-rate changed?

The 2 ways to do that would be:
1. (NewPriceInPoints-OldPriceInPoints) * CtrSize * New_Kwr.Usd_rate
2. NewInPricePoints * New_Kwr.Usd_rate * CtrSize - OldPriceInPoints * OLD_Kwr.Usd_rate * CtrSize

i.e. I can either calculate the sell-buy price difference in points and then multiply it by the current exchange rate and contract size (in that case it doesn't matter what was the exchange-rate when I bought it).,
Or I can calculate both buy and sell prices first in the foreign currency using the 2 exchange rates at the time of buying and selling and then subtract them, which way is correct?

It's just I noticed that my system has been showing big P&L in 3kb lately, when the contract wasn't moving much, so I looked at the exchange rate, which was moving and I started to suspect that I'm calculating it incorrectly..
 
Kernfusion said:
Hi everybody,
sorry for a very basic futures trading mechanics question, but I just need to confirm it for myself:
Is buying a non-FX futures contract denominated in a foreign currency also a bet on that currency or not (not considering margin money)?

e.g. how should I calculate my P&L if my base currency is USD and I bought 1 3ktb contract a month ago, and since then usd.rkw spot exchange-rate changed?

The 2 ways to do that would be:
1. (NewPriceInPoints-OldPriceInPoints) * CtrSize * New_Kwr.Usd_rate
2. NewInPricePoints * New_Kwr.Usd_rate * CtrSize - OldPriceInPoints * OLD_Kwr.Usd_rate * CtrSize

i.e. I can either calculate the sell-buy price difference in points and then multiply it by the current exchange rate and contract size (in that case it doesn't matter what was the exchange-rate when I bought it).,
Or I can calculate both buy and sell prices first in the foreign currency using the 2 exchange rates at the time of buying and selling and then subtract them, which way is correct?

It's just I noticed that my system has been showing big P&L in 3kb lately, when the contract wasn't moving much, so I looked at the exchange rate, which was moving and I started to suspect that I'm calculating it incorrectly..
More...
It depends on when you place actual currency trades. Suppose you open a 3KTB position. Each day that you hold this position a settlement is executed. Either KRW cash is taken out of your account, or deposited into your account. Now you have 2 choices: either you convert each day the KRW cash back to your USD base currency, or you don't.
If you convert KRW -> USD each day you will be using the exchange rate of each day.
If you don't convert until you close the 3KTB position: Now you have a KRW cash position left over. And can either keep this KRW cash position for a future trade, or convert it. You will be using the exchange at the moment of conversion.

My way of working is to keep the foreign currency cash positions, unless there is substantial debit interest being charged. The forex trades also come with commission being charged, so you'll have to balance these.
 
globalarbtrader said:
I'll revisit this subject in my next monthly blog post, in November.

GAT
More...
Hi GAT, don't know if this correlation question is related. However in your excellent first book (page 294), you provide a rule of thumb adjustment from instrument rets correlation to subsys correlation ~ 0.7. I assume is because you are multiplying the covar matrix by 1/2. Out of interest is this adjustment derived from the effects of vol normalization? In my backtests again there is quite a lot of variation around this adjustment, but I appreciate that this rule of thumb might rely on very large samples.
 
Last edited:
I see, so for example let's say the price of 3ktb contract didn't change at all between when I bought and sold it, but the exchange rate changed, then my P&L is zero, so I wasn't affected by krw.usd rate.
(unless I was holding some krw cash for margin, or probably even if I wasn't, the broker landed me that margin in krw for the trade and took it back when I closed the trade., But still this is only a small exposure, just the amount of margin, not the whole notional value of the contract right ?)
 
Kernfusion said:
I see, so for example let's say the price of 3ktb contract didn't change at all between when I bought and sold it, but the exchange rate changed, then my P&L is zero, so I wasn't affected by krw.usd rate.
More...
Correct, if you did not flatten out your KRW cash position every day while holding the 3KTB position.
Kernfusion said:
the broker landed me that margin in krw for the trade and took it back when I closed the trade.,
More...
The broker does charge debit interest for the margin requirements. This is the confusing part (at least with IB) as this margin requirement is not visible as a cash position. But is "eligible" for debit interest being charged to you by the broker.
 
You must log in or register to reply here.
Back
Top