Econometrics: State-space model forecasting - page 6
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I understand all that. I really need half a bar of 100% prediction to make a grail.
Exactly. That's the thing. ))Once in the tester zero bar was available from another symbol. Only symbols selected in the tester are modelled there, and the generated bar for other symbols is available. It was enough for obtaining the tester grail. Watching the euras, trading on the pound.
The variance is nothing. Prediction accuracy is everything. ))
In a textbook. Or in a picture. I haven't seen it in real life.
A TS that (roughly speaking) shows the colour of the next bar (aka trend) and gives a positive mathematical expectation taking into account the trading spread - a myth
You guys are interesting econometricians ... as if you fell from the moon, you are discussing something in your small talk.
If the current price is 1.3200 and the one-step forecast is 1.3200, it is obvious that nothing needs to be done.
The size of the bar exceeds the spread, that's why the forecast of the bar direction is enough. In the first picture almost all directions are correctly predicted, it's fantastic.
may be interesting, but they are realists
see above.
The only thing that is clear to me is that it will not be possible to use the forecast for nothing.
Start small.
Make and test this system on real quotes with no spread at all. If it doesn't work, throw it away...
But if it works (even with drawdowns) - it makes sense to keep digging, because there is something to improve.
If we do not know what to do with it, we will keep on producing errors in our and other people's heads. With clever semantics on the subject of "which is more important - the forecast or variance.... " )))
--
Here's a picture I liked on the fifth forum:
You are wrong to scold econometrics - you live off experience, and if you leave the market for a year, you are left with nothing.
They build the car first, test all the parts, and then they test it on the road. Your tester is a running test. Now it's about testing all the assemblies.I'm trying to find an answer to the question: what to do with the prediction? You are suggesting: TC = model, and there is no such thing, maybe something needs to be done, not justify discarding the model with a tester.
Start small.
If it does not work, throw it away. If it doesn't work, throw it away.
But if it works (even with drawdowns) - it makes sense to keep digging, because there is something to improve.
And until you do it - you will create glitches in your own and other people's heads. With clever semantics on the subject of "which is more important - the forecast or variance.... " )))
--
Here's a picture I liked on the fifth forum:
You are wrong to scold econometrics - you live off experience, and if you leave the market for a year, you are left with nothing.
They build the car first, test all the parts, and then they test it on the road. Your tester is a running test. Now it's about testing all the assemblies.I'm trying to find an answer to the question: what to do with the prediction? You are suggesting: TC = model, and there is no such thing, maybe something needs to be done, not justify throwing out a model with a tester.
and she's not being scolded.
simply applying her methods won't do anything. Zra, do you think faa vandalised the smoking room? Out of desperation.
and she's not being scolded.
simply applying her methods won't do anything. Zra, do you think faa vandalised the smoking room? Out of desperation.
We are different people. Faa posted the text of models that don't work (in my opinion the model doesn't fit the market) and started exploring them in detail - perfect for learning.
I have a model that matches the verbal description of the market: dynamic, for non-stationary processes, adapting in parameters and set of models. But I don't understand how to use it.
In a textbook. Or in a picture. I haven't seen it in real life.
A TS that (roughly speaking) shows the colour of the next bar (aka trend) and gives a positive mathematical expectation taking into account the trading spread is a myth.
It's times like this when you want to hold the writer financially responsible for the bazooka. A bet, maybe?
How much will you lose / win $ if I pay you this "myth" in the tester (on all instruments, for many mullions) in a day? // Timeframe, for simplicity - also a day.
;)
You are wrong to scold econometrics - you live off experience, and if you leave the market for a year, you are left with nothing.
They build the car first, test all the parts, and then they test it on the road. Your tester is a running test. Now we are talking about testing all the assemblies.I am trying to find an answer to the question: what to do with the prediction? You are suggesting: TC = model, and there is no such thing, maybe something needs to be finalised rather than justify throwing out a model with a tester.