Is any prediction doomed? - page 19
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
And what will it give? Isn't it the very probability that many here yearn to predict?
Naturally, it will give the probability of being hit by a car at the crossing and outside the crossing. And, as it should be, it will turn out that at crossings there's a slightly lower chance of getting run over. But, as well as on FX, besides availability of zebra, one has to look around, what not to meet the force majeure :)
PS
That was in response to Paukos' "Most pedestrian collisions statistically occur at pedestrian crossings".
Paukos knows where he's cheating, and unsettled minds can start walking in unintended places.
PPS
Well on the subject: it's not forecasting that's doomed, it's predicting.
I have better things to do.
It seems to me that you are simply unwilling to admit your mistake... I don't think I need to change your mind about anything.
Talking about forecasting, under some far-fetched and imposed conditions -- "first do the math, and then we'll talk" -- is bizarre, to say the least...
If you don't know the subject, I can suggest useful literature for you to read.
I mean, all the aplomb associated with probability and prediction in a simple task that billions of people solve every day, it went into this?
Naturally, it will give the probability of being hit by a car at the crossing and outside the crossing. And, as it should be, it will turn out that the chance of being run over is somewhat lower at crossings. But, as well as on FX, one has to look around not to encounter a force majeure :)
In other words, are you suggesting taking one number, say the number of observations of being hit in a crossing and dividing by another number, which one?
And can the result be considered a probability?
By the way, what are the chances to be run over on crossing, in places where there are no crossings at all? Will this circumstance be reflected in these figures?
Well on the subject: it's not prediction that's doomed, it's prediction.
Instead of words, try you too, to calculate this probability of your beloved successfully crossing the street on a regulated pedestrian crossing. ;) then we can talk about forecasting. It should be much easier than counting probabilities in the market, shouldn't it?
3*18 again! You have a twisted understanding of theorists, even though you are a former mathematician.
Of course it is easier to calculate the probabilities in the financial markets than it is to calculate the probability of yourself going through a transition!
You can run your TS an infinite number of times on the tester using past data as far back as you like and get probability estimates!
But your beloved self how? Run it through the crossroads enough times for statistics? Of course you can, but you must realise that at least one unfavourable outcome in the experiment and it will have to be terminated forever...
Will there be an example of probability calculation in the financial market?
SZY: even if the probability of a price forecast is 50/50 it will already be a "profitable martin".
In other words, you are proposing to take one figure, let's say the number of observations of being hit in a crossing and divide by another figure, which one?
And can the result be considered a probability?
When prediction becomes essentially the same as prediction, it is also doomed.1) Divide by the total number (successful + unsuccessful) of crossings made by pedestrians.
2) You can. Another question to ask here is whether this figure can be used to predict hit-and-runs.
3) If one treats prediction in the same way as prediction, then of course it is doomed.
3*18 again! You have a twisted understanding of the theoretician, even though you are a former mathematician.
Objection, it was said "a little bit of mathematics", that means just one small section of it. Specifically, applied mathematical statistics and modelling of complex systems.
Of course it is easier to calculate the probabilities in the financial markets than to calculate the probability of yourself crossing over!
Would you agree that if you manage to calculate for transition, you can try for the market? And if you fail, you should not try it for the market.
You may run your TS an infinite number of times with the tester on the past data as far back as you want and get estimates of probabilities!
Probabilities of what? Tester grails? What does that have to do with reality.
What about your own self-love? Run it through the crossroads enough times for statistics? Of course you can, but you must realise that at least one unfavourable outcome in the experiment and it will have to be interrupted forever...
Will there be an example of probability calculation in the financial market?
SZZ: even if the probability of predicting a price move is 50/50 it is already a "profitable martin".
already gave an example
1) Divide by the total number (successful + unsuccessful) of pedestrian crossings.
2) You can. Another question to ask here is whether this value can be used to predict hit-and-runs.
3) If you treat prediction the same way as prediction, then of course it is doomed.
1) So divide one by the other, but what's the point?
2) You can't. And rightly so, it's kind of weird to use that value for predicting hit-and-runs. Especially when you're talking about your own carcass.
3) That's what we're talking about. If you are dumbly dividing the amount of something by something, at a crossroads, or in the market, than that is not prediction, which everyone insists on calling prediction.
Objection, it was said "a little bit of mathematics", this means just one small section of it. In particular, applied mathematical statistics and modelling of complex systems.
Would you agree that if you can calculate for transition, you could try for the market? And if it fails, then it is not worth trying for the market. It is worth trying.
Probabilities of what? Tester grails? You can try them too. The probability of positive profit of any TS for a year on any instrument is also possible. Anything is possible. Universal tool, my ass.
That last sentence is a good one. No, it's not. Here we come to what Taleb was talking about. But not to the black swans themselves, which one usually associates with Taleb, but to the part of his book where he tries to impress upon us the value of alternatives. Shall I tell you about the Monte Carlo method?