Pair trading and multicurrency arbitrage. The showdown. - page 320

New comment
 
Sergey2134 #:

I’ve also given a lot of thought to this ‘spread percentage’ parameter – how far apart the instruments have drifted. To measure the spread, you need to have a common starting coordinate for the two instruments.

But in reality, these are two independent charts with different prices. And all we can measure is how far an instrument has moved up or down from its own level, on its own timeframe.

It’s an optical illusion. Isn’t that obvious?

Shift the starting point of the calculation and you’ll get a variety of divergences

 
Renat Akhtyamov #:

It’s an optical illusion. Isn’t that obvious?

Shift the starting point of the calculation and you’ll get a variety of results

< deleted... >

There’s no point in discussing it.

 
Vitaly Muzichenko #:

< deleted... >

There’s no point in discussing it.

If you don’t want to, that’s up to you.
 
That is precisely what we need to discuss. Because, judging by the forum’s history, clever people have already written an indicator showing the spread, and they’ve also written the Tandem expert advisor. And they’ve written a great deal; all of it required further refinement.
 
Sergey2134 #:
Tandem expert

That’s what we’ve been up to, and it’s never worked, which is why we end up with this sort of criticism. I don’t even want to discuss it – there are grown men talking here, and I’m not old enough yet to butt in on their conversation.


 

Hmm… That’s why those threads have been abandoned; there’s no progress.

But what interests me right now is the topic of ‘divergence between two instruments’. If currency pairs are strongly correlated, why not assess the degree of divergence to identify an entry point?

 
Sergey2134 #:

Hmm… That’s why those threads have been abandoned; there’s no progress.

But I’m currently interested in the topic of ‘divergence between two instruments’. If currency pairs show a strong correlation, why not assess the degree of divergence to identify an entry point?

There’s a divergence percentage relative to historical data.
 
Vitaly Muzichenko #:
There is a spread percentage relative to the historical data.
I agree. Right up until your post, I thought the best way to assess the spread was using Bollinger Bands. But over what time period... That’s something to experiment with.
 
Vitaly Muzichenko #:

The most labour-intensive part of the system is determining the movement of pairs, which is exactly what I’ve been working on over the last few days in an attempt to automate it somehow.

I’m somewhat familiar with the subject and can work out how currencies move based on their composition, but since adding the euro and the pound – which weren’t included before – I sometimes find the movements baffling myself: “direct/mirrored”

Is everything else in your system 100 per cent automated?
 
Sergey2134 #:
Is everything else in your system 100 per cent automated?

I suppose so.

I’m currently working on an upgrade; I hope to get a good result

I’m trying to replace the main one with a more informative one.

The bottom two


1...313314315316317318319320321322323
New comment