Wall Observer BTC/USD - Bitcoin price movement tracking & discussion

[European Central Bank]

$1200 happened because Mt.Gox was hacked.. there was trading bot later discovered that kept buying BTC at whatever price, boosting price to impossible range. So hacker has stolen BTC and side-effect was very inflated price.

it was a factor but gox was nowhere compared to china at the height of the bubble. it was already clearly a failing shithole long before it actually died. real people paid real money at those prices.

[Chainsaw]

If you assume price = constant + superimposed sinusoidal curve, with the amplitude of the curve big enough to trigger buying and selling, then by your method, you repeatedly buy low, sell high, over and over again. Which means you make money, assuming the spread and transaction costs are less than what you earn from buying low and selling high.

Any curve can be represented as a sum of sinusoidal curves, i.e. a Fourier series. Therefore, it becomes mathematically provable that your method, if properly implemented, will cause you to benefit from volatility.

EDIT:
Actually I might also have to assume that you start and end at the same price. Obviously if the price of bitcoin shoots up to $1M, you're better off if you didn't sell any of it at all during the rise, which means Strategy 1 would be better.

 All trading strategies have a threshold where they stop being profitable. The rebuy approach works best when ranging. It works worst when on a uptrend with no corrections.
 Even if there is retracing, that retracing must match your thresholds. In the worst-case event the corrections can fall just short of your rebuy levels.

 I built a spreadsheet to play around with formalizing these parameters a few years back.
EDIT - Old version:
https://docs.google.com/spreadsheets/d/1jv97ERhahE7pP5xAVIXtHEE89Ew4CHiz7imtwsYw8zg/edit#gid=4 (Make a private copy before entering your own values.)

New version:
https://docs.google.com/spreadsheets/d/1JDYALoV4KR_pvX5vuQww99t4hwqqmuHuAI9CZWhFgt0/edit#gid=0

In the end you are given a list of several points based on math. The variables you control:

 Net worth outside of bit coin, net worth in bitcoin, desired percentage of assets to hold in bit coin,  granularity of sell targets,  percent of funds to use for repurchases, percent correction rebuy target.

 As an example:  John Doe holds 10 bit coins, as a value of $5000 outside of bit coin, wishes to have 66% of his value in bit coin. Sells occur each 20% increase.  Revise occur when each cell target has dropped in value by half.

 This plan would start out balanced as $10,000 in bitcoin value, 5000 outside, matches the 66% balance. As bitch queen goes up you will be forced to sell more to stay to that balance level. Let's say we jump up to $10,000. At that point you would sell your $10,000 target. Now  in the future, that specific cell will only ever be repurchased if you hit $5000 or lower.

This is the key to the rebuy principle. You don't know how deep the corrections will go.
 If it ever only corrected to 6000, you would never revise anything. If it would have corrected to 2500, you could have three but twice as many. You can't to know in advance. So you have to guess on depth.

 Now, the depth of corrections varies in a rally based on how overextended it is.
So I played around with a "heat index " that let the amount that you sold at price targets increase as you deviate farther from moving averages. This lets you sell more when you believe you are overextended, but again there are no guarantees.

 It's a little cryptic to use with no explanation, and it did not have a ton of interest back in the day. I can find a link to the thread that describes it if anyone is interested now.

 Regardless, it is an important topic, and one that is good to have a good plan for, before prices go crazy and emotions can take over. Also crucial is the tax considerations of your buys and sells, long versus short term capital gains, but I am neither a lawyer nor an accountant so that stuff and all formal recommendations are each individual's responsibility.

[spooderman]

"Sell on news" beware guys.

Also new bitcoin client released, should boost price.

new client boosting price? haha that's a good one. 

[njcarlos]

Now, the depth of corrections varies in a rally based on how overextended it is.
So I played around with a "heat index " that let the amount that you sold at price targets increase as you deviate farther from moving averages. This lets you sell more when you believe you are overextended, but again there are no guarantees.

 It's a little cryptic to use with no explanation, and it did not have a ton of interest back in the day. I can find a link to the thread that describes it if anyone is interested now.

 Regardless, it is an important topic, and one that is good to have a good plan for, before prices go crazy and emotions can take over. Also crucial is the tax considerations of your buys and sells, long versus short term capital gains, but I am neither a lawyer nor an accountant so that stuff and all formal recommendations are each individual's responsibility.

In that vein, what do you speculate is the max depth of any coming correction to this rally?

[Chainsaw]

Now, the depth of corrections varies in a rally based on how overextended it is.
So I played around with a "heat index " that let the amount that you sold at price targets increase as you deviate farther from moving averages. This lets you sell more when you believe you are overextended, but again there are no guarantees.

 It's a little cryptic to use with no explanation, and it did not have a ton of interest back in the day. I can find a link to the thread that describes it if anyone is interested now.

 Regardless, it is an important topic, and one that is good to have a good plan for, before prices go crazy and emotions can take over. Also crucial is the tax considerations of your buys and sells, long versus short term capital gains, but I am neither a lawyer nor an accountant so that stuff and all formal recommendations are each individual's responsibility.

In that vein, what do you speculate is the max depth of any coming correction to this rally?

Not to be cryptic, but that depends on the height of the rally.
The greater the drift from the long term moving averages, the more overbought we are, and the greater potential for depth-of-correction.

If you look at the 2013 rally, we had a highest price over $1000, and a lowest price nearly a year later on one exchange of $100 (and $166 across all).
So even conservatively, we had a $1000 --> $166 = 6x depreciation of the highest prices.

But remember, in a phased sell, you sold coins at 500, 600, 720, etc along the way. So let's say you had your rebuy set at 50% for the 2013 rally.
Any sells you made at 166*2 = $332 would get rebought.
The sells you made below $332 would never be repurchased.

So following the sheet patiently over the 3 year correction would have allowed for a very large profit.
If we instead had stayed flatter, or the bottom held at $500, then the ability for rebuy would have been too 'greedy' in this case, and would have done better with a rebuy at 75% of original value.
Conversely, if we had corrected down to $10 instead of $100, then a rebuy of 25% of original value would have been greater.

There's no way to know the future, so I set those variables based on game theory principles - at what values will I be okay, agnostic of the path Bitcoin takes?

There was a case study of mis-gauging this burning an individual from our most recent rally. IIRC, masterluc got burned during the rally to $500 in 2015. In short, it got 'overextended' too much so he sold ABOVE the planned amount. As soon as you do that, your choices go away. If price keeps going up, you cannot benefit. So we had already been a few standard deviations above expected maximum price...and kept going.

For me, that's the biggest benefit to the spreadsheet. I can be greedy by deferring some sales to later price points, but I never sell more than the limits set BY those price points.

...my best speculation is that in the long term, this rally peaks at $8,888, and corrects back to between $500 and $1000.

[njcarlos]

For me, that's the biggest benefit to the spreadsheet. I can be greedy by deferring some sales to later price points, but I never sell more than the limits set BY those price points.

...my best speculation is that in the long term, this rally peaks at $8,888, and corrects back to between $500 and $1000.

How can we learn more about this

[JimboToronto]

A late good morning Bitcoinland.

I see we had a nice little correction last night but not all the way back to triple digits, then bounced back up to a new yearly high before settling down to where we are now... $1029USD at Bitcoinaverage.

You gotta love the corrections happening as we go along instead of letting them pile up and bringing a crashing end to the rally.

I love the daily CAD ATHs... $1393 today before dipping back to the current $1382.

Go Bitcoin go.

[BTCtrader71]

 I built a spreadsheet to play around with formalizing these parameters a few years back.
https://docs.google.com/spreadsheets/d/1jv97ERhahE7pP5xAVIXtHEE89Ew4CHiz7imtwsYw8zg/edit#gid=4 (Make a private copy before entering your own values.)

Cool. I played around with a similar google spreadsheet a while back. Yours looks a whole lot more polished than mine

Mine started out as my own modification of the Sane and Simple Bitcoin Savings Plan. Conceptually, right now, my idea would be to create a bot that does the following: whenever X bitcoin is sold at price P for D dollars, create a list of buy orders spread out over a wide price range which, when totaled, would use the entirety of the D dollars to buy back X + a little extra bitcoin. Example: if the algorithm sells 1 btc at $1000, it automatically places buy orders for $200 worth of bitcoin at prices $900, $800, $700, $600, and $500. If the price falls all the way to $500, then you will have used up the entire $1000 to buy back bitcoin at a cheaper price. You can make it as granular as you think the bot and the system can handle. The interesting problem becomes: what is the ideal way to distribute the buybacks? The answer should be determined by what you expect the volatility to be, so that you can maximize how much you profit from volatility. If there is zero volatility (steady rise to the moon), you're basically just implementing the SSS.

One other thing: whenever you buy bitcoin, your bot should do the same thing in reverse, i.e. it should create a spectrum of price points where you sell it all back, but at higher prices. From an algorithmic standpoint, just flip the BTC/USD pair and treat buying btc as if you are selling USD.

[Chainsaw]

For me, that's the biggest benefit to the spreadsheet. I can be greedy by deferring some sales to later price points, but I never sell more than the limits set BY those price points.

...my best speculation is that in the long term, this rally peaks at $8,888, and corrects back to between $500 and $1000.

How can we learn more about this

EDIT: Old version of sheet:
I linked to it in the first post, a few messages back. Here's that link: https://docs.google.com/spreadsheets/d/1jv97ERhahE7pP5xAVIXtHEE89Ew4CHiz7imtwsYw8zg/edit#gid=4

New version:

https://docs.google.com/spreadsheets/d/1JDYALoV4KR_pvX5vuQww99t4hwqqmuHuAI9CZWhFgt0/edit#gid=0

And here is the original thread for discussion around its usage (also linked from profile signature):

https://bitcointalk.org/index.php?topic=326308.60

[njcarlos]

I linked to it in the first post, a few messages back. Here's that link: https://docs.google.com/spreadsheets/d/1jv97ERhahE7pP5xAVIXtHEE89Ew4CHiz7imtwsYw8zg/edit#gid=4

And here is the original thread for discussion around its usage (also linked from profile signature):

https://bitcointalk.org/index.php?topic=326308.60

Thank you! I'll check it out now.

[Chainsaw]

I just made a public copy of my most recent version.
The old one I had linked still used a few more awkward data entries, the new one linked to is as described above, I like it a bit better.
Both exist, so you could use either.
I'll update some of the other links to the new one so you can play with whatever works best.

The new version: https://docs.google.com/spreadsheets/d/1JDYALoV4KR_pvX5vuQww99t4hwqqmuHuAI9CZWhFgt0/edit#gid=0

[jbreher]

Any curve can be represented as a sum of sinusoidal curves, i.e. a Fourier series.

Not quite. Any periodic curve can be represented as a sum of sinusoidal curves. The trick is in knowing the periodicity of $=f(t). protip: that f() ain't periodic.

Nope.
An infinite fourier series can match any continuous differentiable function.

An infinite series of a non-deterministic function is unrealizable.

Unless you know the function a priori, you cannot derive a convolution thereof. If you know the function a priori, there is no need to convolve it in order to profit. The entire concept of obtaining a Fourier series of $=f(t) is therefore nonsensical.

[BTCtrader71]

Any curve can be represented as a sum of sinusoidal curves, i.e. a Fourier series.

Not quite. Any periodic curve can be represented as a sum of sinusoidal curves. The trick is in knowing the periodicity of $=f(t). protip: that f() ain't periodic.

Nope.
An infinite fourier series can match any continuous differentiable function.

An infinite series of a non-deterministic function is unrealizable.

Unless you know the function a priori, you cannot derive a convolution thereof. If you know the function a priori, there is no need to convolve it in order to profit. The entire concept of obtaining a Fourier series of $=f(t) is therefore nonsensical.

Obviously we don't know what the price curve is going to be in the future. That's not the point. The point is that if the price starts at f(t_0) = X and ends at the same price f(t_1) = X, and in between it has lots of ups and downs, then this method will increase the end-value of your total holdings (expressed in BTC or in $). It doesn't matter what f(t) does in between. You don't have to know ahead of time what f(t) is going to do. The demonstration of this statement makes use of the fact that f(t) = the sum of lots of sine waves, for any f(t). This is true even if we don't know ahead of time what exactly f(t) is.

[Soros Shorts]

Any curve can be represented as a sum of sinusoidal curves, i.e. a Fourier series.

Not quite. Any periodic curve can be represented as a sum of sinusoidal curves. The trick is in knowing the periodicity of $=f(t). protip: that f() ain't periodic.

Nope.
An infinite fourier series can match any continuous differentiable function.

An infinite series of a non-deterministic function is unrealizable.

Unless you know the function a priori, you cannot derive a convolution thereof. If you know the function a priori, there is no need to convolve it in order to profit. The entire concept of obtaining a Fourier series of $=f(t) is therefore nonsensical.

Obviously we don't know what the price curve is going to be in the future. That's not the point. The point is that if the price starts at f(t_0) = X and ends at the same price f(t_1) = X, and in between it has lots of ups and downs, then this method will increase the end-value of your total holdings (expressed in BTC or in $). It doesn't matter what f(t) does in between. You don't have to know ahead of time what f(t) is going to do. The demonstration of this statement makes use of the fact that f(t) = the sum of lots of sine waves, for any f(t). This is true even if we don't know ahead of time what exactly f(t) is.

There is a simple way to do a sanity check to determine if what is proposed is nonsensical.

1) Find a price level X where the curve hits at 2 points, t_0 and t_1. Solve for f(t_0) = X and f(t_1) = X.

2) Find another price level Y where the curve also hits at 2 points, t_2 and t3. Solve for f(t_2) = Y and f(t_3) = Y.

Use the 2 solutions to extrapolate 2 curves into the future. If the curves are wildly different then all this would be a waste of time.

[wutizurkwest]

New forecast from masterluc

https://www.tradingview.com/chart/BTCUSD/fvhCAiH9-EW-development/

[arklan]

i've been away for months dealing with, you know, life. but i've kept an eye on the price.

seems like a steady climb this time, not any sort of real sudden climb. maybe we'll hold it this time.

[keithers]

It's crazy how that if BTC were any other "traditional" investment, it would be front page news of any news outlet because of its rapid appreciation. I guess we'll need to wait for another hack or some sort of FUD for BTC to be the main headline again

[rjclarke2000]

Looks like under 1000 is coming. What is everyone else seeing over the next few hours?

[K~Ehleyr]

Looks like we're having another little correction before the next flight upwards. Looking good as far as I can see, consolidating as we go along!

[Lauda]

Who is leading this 'dump'? It would be nice to see 1000$ hold.