Fully automated futures trading

Kernfusion said:
Seems I'm always starting the sad conversation :), but did we just loose some money? I'm down about 5.5% in the last week or so (lost on UB, SEK, NZD, MHI), also the system dropped a lot of exposure., Right after things started to look up..
More...

Yes from a mini HWM (actually about 1% from my absolute HWM) on the 1st, I'm down 5% at pixel time. Bonds as you'd expect but also FX and equities.

Code: 
********************************************************************************
Profit and loss report produced on 06/11/2023 08:50 from 01/11/2023 13:00 to 06/11/2023 08:50
********************************************************************************


Total p&l is -4.589%

====================================
P&L by instrument for all strategies
====================================

           codes  pandl
0           US20   -1.0
1            DAX   -0.4
2        RUSSELL   -0.4
3           BUND   -0.4
4            AUD   -0.4
5            NZD   -0.3
6            CHF   -0.3
7   NASDAQ_micro   -0.3
8            GBP   -0.3
9            JPY   -0.3
10       EURO600   -0.1
11        EU-OIL   -0.1
12  COPPER-micro   -0.0
13       BITCOIN   -0.0
14       LEANHOG    0.1
15        NIKKEI    0.1
16           MXP    0.4


==========================
     P&L by strategy    
==========================

              codes  pandl
0  dynamic_TF_carry  -3.72
0          residual  -0.87


==================
P&L by asset class
==================

    codes  pandl
0    Bond  -1.38
1      FX  -1.22
2  Equity  -1.12
3  Sector  -0.07
4  Metals  -0.02
5     Ags   0.09

These savage end of market turning points are always going to be painful. Better to be a discretionary trader perhaps:

https://twitter.com/x/status/1720078981045903507



Rob
 
Hi Rob / All,

I’m using CFD per point for my calculations for position size, my broker is CMC Markets here in the uk, but something I don’t get is why with Orange juice do I need to multiply the answer by 100 to get my units but with Cocoa I don’t?

For example -

Orange Juice :

Capital £4000

Annual std dev - 30.56%

Price - $350



Notional Exposure = (12% x 4000) / 30.56% = £1570

CFD Per Point = (£1570 x 1.24 x 1) / $350 = 5.56


But on my brokers order ticket the minimum units I can trade is 300 for OJ, but if I multiply 5.56 by 100 it works out almost bang on to my notional exposure, whereas with Cocoa (see below) I don’t need to multiply it by 100. Why is that?


Cocoa :

Capital = £4000

Annual std dev = 19.40%

Price $3882



Notional exposure = (12% x £4000) / 19.40% = £2474

CFD Per Point = (£2474 x 1.24 x 1) / $3882 = 0.79



See pics attached.

OJ and Cocoa.png

OJ-and-Cocoa.png

vjG7D5w


Thanks for any help,

Tom
 
Last edited:
peppermint_tea said:
Hi Rob / All,

I’m using CFD per point for my calculations for position size, my broker is CMC Markets here in the uk, but something I don’t get is why with Orange juice do I need to multiply the answer by 100 to get my units but with Cocoa I don’t?

For example -

Orange Juice :

Capital £4000

Annual std dev - 30.56%

Price - $350



Notional Exposure = (12% x 4000) / 30.56% = £1570

CFD Per Point = (£1570 x 1.24 x 1) / $350 = 5.56


But on my brokers order ticket the minimum units I can trade is 300 for OJ, but if I multiply 5.56 by 100 it works out almost bang on to my notional exposure, whereas with Cocoa (see below) I don’t need to multiply it by 100. Why is that?


Cocoa :

Capital = £4000

Annual std dev = 19.40%

Price $3882



Notional exposure = (12% x £4000) / 19.40% = £2474

CFD Per Point = (£2474 x 1.24 x 1) / $3882 = 0.79



See pics attached.

View attachment 326856
OJ-and-Cocoa.png

vjG7D5w


Thanks for any help,

Tom
More...

This happens with futures as well, but with some agricultural instruments you need to multiply the price shown by some factor to get the actual price. Basically a quoting covention.

Rob
 
I was down about 5% last week, mostly due to short positions in US treasuries/equities and long USD. I always get nervous when there is a Fed meeting since they sometimes to lead to large changes in equities, rates, and currencies.
 
globalarbtrader said:
1. It's obviously the r* in the formula immediately above on page 79; and you would use the same lambda as when calculating the standard deviation
2. Why would you think that say Gold would go up by more than any other asset? That's not just overfitting, it's discretionary trading...

Rob
More...

I want to express my thanks for your helpful explanation. Now, my representation of N closely follows the one shown in Figure 16 on page 81, thanks to your guidance. I really appreciate it!

You're right about the strategy for Gold. I should concentrate on a designed strategy rather than a tailored one. I've been thinking in a discretionary way for so long that I realize there's still a long way to go!

I won't ask you to explain all the book details and examples, but I do think it's crucial to establish a strong foundational understanding before moving forward. The annualized standard deviation of percentage returns seems vital in all the strategies you've presented. I hope others reading this will benefit from your insights too.

I've got a few questions about Figure 18 on page 89 and cryptocurrency.

  1. Why use a two-month rolling σ% instead of one month? In a previous section (p.78), you mentioned the effectiveness of the last month's standard deviation in forecasting future volatility.
It seems I might have misunderstood something about "rolling" or made an error in the formula. I initially calculated the 'returns percentage,' then calculated standard deviation of returns (with a window size of 40 days), and annualized it by multiplying by 16. However, the result seems different from what I expected.


2. Do you find no interest in trading alt coins (cryptocurrencies) besides BTC or ETH?
I found out that they are very volatile but comparatively easier to manage the contract size.
but they definitely lack the history...​


Your help in clearing this up would be much appreciated.

Thanks again for your time and guidance.
 
No.Emotions.Trader said:
I've got a few questions about Figure 18 on page 89 and cryptocurrency.

  1. Why use a two-month rolling σ% instead of one month? In a previous section (p.78), you mentioned the effectiveness of the last month's standard deviation in forecasting future volatility.
It seems I might have misunderstood something about "rolling" or made an error in the formula. I initially calculated the 'returns percentage,' then calculated standard deviation of returns (with a window size of 40 days), and annualized it by multiplying by 16. However, the result seems different from what I expected.


2. Do you find no interest in trading alt coins (cryptocurrencies) besides BTC or ETH?
I found out that they are very volatile but comparatively easier to manage the contract size.
but they definitely lack the history...​



Thanks again for your time and guidance.
More...

1. Because I'm trying to show a completely different point, which is ex-post vol not vol forecasting, and if I used one month the graph would be too noisy to interpret.

2. Because I would have to own spot crypto in my own wallet, or be exposed to a dodgy crypto currency exchange (And they are all dodgy, sorry). Spot crypto would be a huge amount of work to set up properly (I've tried and failed to buy spot crypto in the past, as an experiment). More importantly I hate crypto with a passion, it is a planet destroying, wealth destroying, time wasting ponzi scheme.

My exposure to crypto is limited to trading them with futures (where they could be just about anything frankly), and taking my 30 silver pieces on a regular basis from a crypto hedge fund that I do consultancy for. So if the alt coins were listed on the CME, I'd trade happily trade them.

Rob
 
Hi guys, I'm setting up a new system and have trouble in creating multiple prices. Can't get prices for USIRS5ERIS as it is removed but then creating multiple prices fails on this instruments. What is the proper way to remove USIRS5ERIS?
 
globalarbtrader said:
This happens with futures as well, but with some agricultural instruments you need to multiply the price shown by some factor to get the actual price. Basically a quoting covention.

Rob
More...

Just wanted to say if this helps anyone, i found that because some quotes are given in cents, if you convert it from cents to dollars you get the correct number of units without needing to multiply at the end. Either method obviously works, but it was good for me to understand.

Eg OJ is quoted in cents with my broker, say 374.78 - so if i just put $3.75 in the formula i get the right number of units straight off the bat without needing to multiply the final answer.

I'm sure this is rookie stuff but great to understand for me personally.

(I also learned that USX means cents on TradingView)

Thanks
 
You must log in or register to reply here.
Back
Top