Topic: : - page 19. (Read 70881 times)

Do you think analyzing past changes in bid/ask depth (in addition to past change in price) could result in more accurate price predictions?

Make some predictions and statistically compare them with the theory that the price will remain constant.  Sum the square of the variance (predicted value - actual value) and see what does better.  Do this 20-100 times and you'll have a good idea of how your model performs.

Chodpaba,

Just wanted to express my thanks for your continued analysis. I enjoy these graphs and projections. I don't have much to share myself, the model I was working on hasn't been of much use lately.

What about a randomly irrational person?

Predictable? 

Ahh, interesting!

of course you can predict irrationality, don't you have any stupid friends that are predictably irrational?

you're think of another word or something, maybe "unlikely" or "improbable" thinking if you can see it coming it's not improbable anymore, but irrational isn't the same, people can be predictably irrational and that's nothing to do with this market 

I'm not sure you CAN capture irrationality.
It is per definition not a pattern (well, maybe there is a meta-pattern there?)
It can be captured if there are only two choices, one rational and one irrational.
Then you can negate it, irrational = NOT rational.
But i don't think there is this type of choise.


So i'm wondering, aren't your predictions based on bots and people that exhibit predictable behavior?

so back down to 8/9.5 by Thursday?  ,  I'm pretty sure that's past a lot of people's stop losses now, I think if were back at 8 were going even lower.

Nothing against your efforts, but I really do not think you can use this model to predict financial markets. Financial markets including bitcoins work irrationally and based upon mass psychology. This analysis is very complex and require the use of other instruments, like technical analysis, indicators, sentiment, wave patterns, etc. on top of mathematical models.

One example are the small ranges you give every day. This works for a while, but then there is a huge breakout which you did not predict. For example in the technical analysis  we saw that a break below 12.5-12.9 will quickly lead to 10$ then 5-5.82 $ and then a subsequent rise to at least 10$, which happened almost exactly.

But you did not predict those up to 100% moves.

All the indications are that we are headed for a severe financial crisis, as that realization dawns on more and more people, they are going to be looking out for security, how can I protect my savings, how can I buy food, Silver Gold etc.
Bit coin will only succeed if we can draw in the masses, this means they must see businesses up and running. The masses are not interested in speculation, they have lives to lead, time to spend with their family, all they want is a quick easy way to buy stuff, silver, food, a cinema ticket, a gift for a loved one.
If they can do that AND stick two fingers up to the establishment, they will buy in.
If those things are available, they will buy into Bitcoin, and the more that buy in, the more the price will increase, it's that which will drive the price stratospheric, online vendors using bitcoins.
Bitcoiners can help ensure the success of Bitcoin by building businesses that accept bitcoins and by making purchases with Bitcoin.
I know there are many already but the success of bitcoin depends on us all taking action that supports bitcoin, that means settingup businesses and / or purchasing with bitcoin.
Those businesses already set up, thse people who've made the commitment shuld be rewarded, either with our custom or small donatins, it is they that will ensure the success of Bitcoin, not the Miners, and not the hoarders and not the speculators, without bitcoin traders, people who buy and sell using bitcoin, the masses will not buy in.
Those who have made the effort should be encouraged and rewarded for their effort, is there any way that some of the Bit coin fees can be distributed to those who have set up businesses online or who accept bitcoin, this not only rewards their efforts but provides incentive for others to have a go at setting up bitcoin businesses.


 

I don't think he did that bad either, the 120k investor threw a spanner in the works prediction wise, but without that he was pretty close, you can't predict everything in a tiny market there is so many early adoptors have massive amounts of coins if something happens in there life and they need cash ASAP or a large investor suddenly want's in, it will mess up ALL predictions.

and the

2011-08-14   $9.207   $10.130
2011-08-15   $8.808   $9.490
2011-08-16   $8.840   $9.133
2011-08-17   $8.775   $9.295

predictions

I was looking at that earlier wondering if we could manage get up to $10 for your "2011-08-14   $9.207   $10.130" prediction and here we are, so don't give up chodpaba there is definitely something to your type of analysis.

His model showed a possible short term bottom in the 3s, with strong support at 5.80.  It also showed a possibility of $13 as a short term resistance.

Both of these scenarios were realized in a 48 hours from the time of the prediction.  As far as I'm concerned that's just uncanny, and a much better record than my ass which saw $5.80 as well, but not the $12 shortly thereafter.

If his model could make the right prediction in every scenario he would be making tons of money and would not have been here discussing it with us.

I guess the better way to look at the forecast isn't "what will likely happen" but "what will likely happen if trends remain constant in direction and magnitude." To predict group variability with total accuracy is impossible; can you tell me what the overall luck for all miners will be tomorrow? Likewise, predicting outliers isn't realistically going to happen as the events that triggered the crash was a chain event, a butterfly effect, not just one big drop.

i have to wonder if a large event(s) happens like the exchange losing it's btc and mybitcoin finally known to be owned... would
throw out any guess on pricing and cannot really be held against a person....

so here we have a bitcoin conference coming up and i have to wonder if chodpada is putting that into his calculations at all.
we could very well see people buying in expectation that some media attention might create a burst up days/week after the conference.

or is a pure math guess without taking into account external influences.. basically assuming several normal trading days without big news
to look forward to.

i have not read this whole thread.

All due respect (IMO your a very intelligent guy) but didnt your analysis also predict a value of $10 - $13 over the next few days the day before it crashed to $6 

So it's kind of a fuzzy logic approach, where you take all these different factors into account with one huge equation? Like,

new_Price= Old_Price * (1 + 0.1 if it's Friday - 0.05 if Volume was low yesterday + ...)

How are you coming up with these price targets? Do you fit a polynomial to the data?

It's shocking we had as much correlation as we had.

Everyone came to a concensus on a 5.80 as a rock solid resistence using our collective crystal balls, analytics, guts, asses and lo and behold that's exactly what happened.  You also called a relatively immediate $12, which my ass didn't see as likely at all.

So overall, excellent predictive results even though time frames don't seem to match up precisely.  If only we had derivatives which let us act on expected volatility.   Then again, if we did have them perhaps we wouldn't have such spectacular manic/depressive cycles.