Topic: Stephen Reed's Million Dollar Logistic Model - page 16. (Read 123190 times)

Bitcoin market cap this moment - $ 5,615,035,838

# of countries with federally held bitcoin reserves - 0

# of major banks offering bitcoin-related products/services (not counting subsidiary exchanges) - 0

# of trusted organizations offering insured bitcoin storage - 1 : with reasonable parameters - 0

if we fall below trend / stay stagnant for a time to dislodge from trend, does not mean much fundamentally of course.

it may not even hold much against said trend, as a future bubble may 're-align' us etc.

In which case most of us involved with BTC at this point should be in a pretty good place by then.

It does seem like we should be nearing the end of this long and a bit painful consolidation period. 

I'm thinking the model is 'correct so far'.

That is, price is behaving according to the model, but here's the thing, his model could be wrong because the price levels out at 500k or 1.5 million

Model, Shmodel, who needs a model...

Um, who should we take more seriously...  a guy who says he has a model but then won't provide it... aliro38... or this guy...

• I guarantee there will another Bitcoin price bubble... -- Barry Silbert
• Big milestone today for Bitcoin Investment Trust -- now holds over 100,000 bitcoin -- Barry Silbert

I'd say we are right in the slope of enlightenment

I hope it's not the case, otherwise we may have to revise the projections to fit the hype cycle instead. And if that's the case, then perhaps we hit the peak of "inflated expectations" last November.

Here is to hoping SlipperySlope's model is correct.

I would think that even if we continue on for several more weeks at a plateau or even just slight movement upwards we are not too far off of the model though are we? 

That is my appraisal of the current situation too.

Price forecasts are meaningless without a model imho. I can forecast the price for tomorrow with more precision if compared to any multi-year model but that's just blind following the market with reduced error (due to much smaller timeframe). Pretty much useless, if not harmful,  in the long term if you ask me.
How can you spot bitcoin off track without a model? In fact, if my interpretation of SlipperySlope's model is correct, we are either in a "Moment of maximum opportunity" but also "Maximum risk of model breaking" because we are so close to historical boundaries of model itself.

I know my english is far from perfect, apologies for that.

Thanks to all for your great feedback!

Indeed, maybe it’d be much better to start a new thread, so this one stays clean.
My intention is to post weekly forecasts, both for the record and for testing purposes. If they’ll reasonably confirm then we shall move on, or otherwise my work will prove itself useless, wouldn’t it?
Please feel free to come and criticize/comment there.

I don’t have yet a sensitivity/error analysis neither many other things (my work is incipient), but obviously the timeframe is much shorter than that of Logistic Model. But rest assured it is not just a TA on price charts. For those interested in the inner workings, my starting points and ideas were already posted here: https://bitcointalksearch.org/topic/m.6452694

The problem with timeframe, as I see it now, is that bitcoin network seems to correct quite violently in terms of price evolution, both upwards and downwards. Such abrupt corrections may be modeled based on an internal variable with a threshold value. Once the threshold is even slightly exceeded, correction beings to unfold. Bellow that threshold the sensitivity is close to zero, where many parameters like price, transaction volume, number of transactions etc are free to float with little significance on the big picture. The essential feature is that when bitcoin reacts it does (or at least it always did) so toward a quick convergence with the Logistic Model / exponential growth. I’d say this kind of behavior was already felt by most of us but having it put inside a model is not trivial. It’s worth a try, nevertheless.
The big challenge is, imho, to parameterize the convergence and to periodically check if it holds true. If it does, then Logistic Model is valid, shorter timeframe models are pretty useless (maybe extremely useful to traders but that’s all) and for everyone else the good news is bitcoins can be hold on long term as an excellent investment. But if the convergence limits are breached (assuming they were well defined and modeled in the first place) it would mean, as sad as it may sound, that bitcoin is off the track from exponential grow, possibly to switch itself to another growing curve, to stagnation or even to a fade-away.

I’ll soon post the new link and sorry for slightly polluting the great thread here.

Indeed, but a week may be a bit to short a period for events to unfold.

He's not the only one...


https://www.tradingview.com/v/hjgUvEKf/
https://www.tradingview.com/v/zP3USOow/

It's going to be an interesting week for bitcoin...

Well, for starters, he's predicting under $400 by May 31.  It remains to be seen, but that would be interesting if it happens.

I'm genuinely interested, but without a model you just made a price forecast. There's no point comparing it to SlipperySlope's model as the timeframe and historic volatility, among others parameters, must be considered.
If you think the last 12 months followed a different pattern, best suited by your model, you should explain it somehow as it would be of great value for everyone.

Have a nice week too!

You're welcome.The model is on my computer, so far scattered in several files but I'll clean it up.
It is not based on a particular modelling theory. I've started it from scratch, looking for correlations.

Have a nice week.

The logistic model is here . . .

https://docs.google.com/spreadsheet/ccc?key=0ArD8rjI3DD1WdFIzNDFMeEhVSzhwcEVXZDVzdVpGU2c

http://en.wikipedia.org/wiki/Logistic_function

aliro38's model?

Pardon me, but where's the model?

Since my last post here, I’ve managed to develop what seems to be a tremendously strong model. (In fact, it was more like inspiration literally came and hit me...)
There is still work I have to do on automating and slightly refining the calculus (currently I do a lot of hand-on work and I use some shortcuts/simplified/linear math) but I’m truly amazed. The model basically matches the last 12 months with great accuracy (I haven’t computed it yet but by far it looks better than anything I’ve ever seen in stochastic models. In fact, a correlation that good can’t be of stochastic nature and at least its core is deterministic, which made me go 'hmmm…').

The first “hot test” was already launched here, albeit modestly: https://bitcointalksearch.org/topic/m.6618571
Maybe the launch was a little too early, too rough (I didn’t fine tune anything) and too far in time for an accurate prognosis (>20days) but I think it will hold.

Why I’m posting all of these?
Well, in the simplest linear extrapolation I’ve done 4 days ago, the price for today (May 11) in my model was predicted to be 377-442$. For yesterday it was predicted at 382-444$ (a bit under evaluated but again my first forecast was quite crude, linear and some figures were mercilessly rounded), for tomorrow prediction is at 373-439$ and finally for the day after tomorrow it will be at 368-437$. The model is dynamic, meaning that some time-dependant critical variables will have inflexion points that, when happening, will invalidate current prediction but mind it there were only two inflexion points in the last 365 days!
I’ll do a proper (2nd or 3rd order) extrapolation as soon as I’ll have some time in my hands but due to the amount of math involved it may not be happening until next weekend.

Finally, to the points:
1. I confirm the downtrend and I am to determine whether it is accelerating, constant or slowing down toward a possible reversal.
2. I don’t want to contradict the Logistic Model as I believe it to be really great and correct as macro model in long run (or at least until its hypotheses are questioned as no longer being valid) but rather to suggest considering that there are models out there (I have reasons to believe that chances are I might have re-discovered what others may already be knowing/applying for a long time). Such models (mine included) predict the current price is not a bargain but it is the correct one according to the bitcoins’ fundamentals.
3. I believe current price might actually be at its upper margin in the downtrend channel.
4. Of course, all of the above may be seen for now (and proved by future evolution) to be mere under-performing in the Logistic Model but if the discrepancies with Metcalfe’s Law (in terms of number of transactions etc) keeps growing it may be a sign for questioning the exponential growth.
5. Let’s see what happens by May 31, compare the 292-398$ prognosis with that of >1232$ (please update) and see if further comparisons may have some merits. Agree?

Best regards,

Thanks for the updates on the model. Here's a chart using Bitstamp prices from TradingView Charting tool attempting to capture the median line and price action above and below it.

Indeed. SlipperySlope was my then-so-clever handle at the peak oil blog The Oil Drum, which I carried over here back in 2011. On Reddit I am frugal-guy. The SlipperySlope user name here has no semantic relationship to Bitcoin.

-Stephen Reed