TheKoziTwo
Legendary
Offline
Activity: 1552
Merit: 1047
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:08:36 PM |
|
Something is wrong with that chart. I have seen many log charts (for example the one at bitcoincharts.com) but they all look different. The first thing that's wrong about that chart is that it's pulled out of the OP's ass. He was drawing a line based on "Mmmm, right about there". ok then. Plotted using sigmaplot. regression analysis in genstat stats package (version 12) minus bubble data. bitcoin price (logged) = -0.0635+0.0013791*time predicted price= $24.24 when I first saw this, I sold everything. edit: all bitcoins ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) Nice, so you actually didn't pull your prediction out of your ass, but used (polynomial) regression on historical data (not sure if it's wise to exclude bubble data/outliers, though). I respect that. But do you want to know what is the single biggest assumption Dude, you either don't know the math behind what you do, or you're willfully obstinate to make a point: if you run a simple regression analysis like you did, you de facto work from the assumption that there is one and only function underlying those data points. Otherwise, you "identified" jack shit. I'm only showing that there's an extremely strong underlying trend outside of bubble data, which predicts a bitcoin price much lower than it is now. No need to get antsy ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) eh, sorry. my field (well, of sort) is foundations of mathematics, so I can get worked up over not knowing the (formal) premises of a method one uses. anyway, I'm not even completely denying your point. I'm extremely cautious right now, and far from ruling out price will drop further. But I still consider it a distinct possibility that the fundamentals and public perception have changed enough for growth to continue from (roughly) where we are now. Time will tell. ok truce. FYI, I'm a statistical modeller. Biology? Bioinformatics? Behavioural biologist, but don't judge me. Why would I? I'm a mildly formal philosophy person, basically ![Tongue](https://bitcointalk.org/Smileys/default/tongue.gif) EDIT: sorry for the off topic conversation. it's just that nothing happens right now. ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FdlOv6LD.gif&t=665&c=-e6DnbSkFOR7Zw)
|
|
|
|
prof7bit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:09:50 PM |
|
bitcoin price (logged)
Then please explain why your "bitcoin price (logged)" does not look anywhere similar (not even remotely) like any other bitcoin price log chart? It actually does. No: ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fi.imm.io%2F18kQP.png&t=665&c=pTf5k6XQZP1B3g)
|
|
|
|
ElectricMucus
Legendary
Offline
Activity: 1666
Merit: 1057
Marketing manager - GO MP
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:14:19 PM |
|
That chart may actually be the 7day moving average of the price or something
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:17:21 PM |
|
bitcoin price (logged)
Then please explain why your "bitcoin price (logged)" does not look anywhere similar (not even remotely) like any other bitcoin price log chart? It actually does. No: ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fi.imm.io%2F18kQP.png&t=665&c=pTf5k6XQZP1B3g) To linearize things (make them 'straight' for analysis) that have points below and above 1, you add 1 to your log (i.e. log x+1). It looks like they haven't done that here. Simples?
|
|
|
|
phoenix1
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:19:32 PM |
|
Hey guys Would you mind 'snipping' your posts ... no problem with chit chat - as you say, there is nothing going on, but my scroll finger is getting tired ![Tongue](https://bitcointalk.org/Smileys/default/tongue.gif) Glad you settled you differences ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) EDIT: I see a new dispute has arisen ... have fun ![Grin](https://bitcointalk.org/Smileys/default/grin.gif)
|
|
|
|
prof7bit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:21:29 PM |
|
To linearize things (make them 'straight' for analysis) that have points below and above 1, you add 1 to your log (i.e. log x+1).
No you dont! You might have accidentally done log(x+1) because log(x) +1 would not matter, it would only shift the entire graph up on the y axis while your log(x+1) totally distorts it. You could produce even more funny meaningless curves by using log(x+42) or log(x+23) or whatever arbitrary constant you add to the price before logging.
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:22:49 PM |
|
To linearize things (make them 'straight' for analysis) that have points below and above 1, you add 1 to your log (i.e. log x+1).
No you dont! You might have accidentally done log(x+1) because log(x) +1 would not matter, it would only shift the entire graph up on the y axis while your log(x+1) totally distorts it. not when it's logged. try it with numbers above and below 1.
|
|
|
|
Rampion
Legendary
Offline
Activity: 1148
Merit: 1018
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:24:25 PM |
|
I have to say that after the April, 10th crash I placed big bids around $30. Without much analysis, I strongly believed that the post-crash bottom was going to be around $30, which is the 2011 bubble top.
For that to happen I was expecting a 2011 situation in terms of negative news coverage, "this is a ponzi" screaming, etc.
After we rebounded at $50 with two consecutives all time highs in terms of daily volume, I changed my mind. I realized there are many differences compared to 2011. After the last bubble, people thought Bitcoin was dead. News coverage was extremely negative. There were posts on these forums saying Bitcoin was a huge ponzi, with Satoshi as its very own Maddoff.
Things are very different now. I really believe we won't go below $50 unless there are some very bad news. Thus, even if $25-$30 coins didn't seem crazy to me after April, 10th, I now believe it's only a wet dream of those willing to increase their BTC stash - regardless of what TA says.
|
|
|
|
prof7bit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:25:32 PM |
|
not when it's logged. try it with numbers above and below 1.
No. Add parenthesis just to be sure. Your graph is wrong, the one at bitcoincharts is correct.
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:27:57 PM |
|
not when it's logged. try it with numbers above and below 1.
No. Add parenthesis just to be sure. Your graph is wrong, the one at bitcoincharts is correct. why do I feel like that will be your conclusion whatever I say? What does bitcoinchart predict? edit: just realised I misread you - Log (x+1) is what I did, log (x+1) was what I meant to do. It linearizes it like I said, which is one of the main reasons for logging charts.
|
|
|
|
oda.krell
Legendary
Offline
Activity: 1470
Merit: 1007
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:28:26 PM |
|
Hey guys Would you mind 'snipping' your posts ... no problem with chit chat - as you say, there is nothing going on, but my scroll finger is getting tired ![Tongue](https://bitcointalk.org/Smileys/default/tongue.gif) Glad you settled you differences ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) EDIT: I see a new dispute has arisen ... have fun ![Grin](https://bitcointalk.org/Smileys/default/grin.gif) Yes, sorry, completely forgot about that :/
|
|
|
|
ElectricMucus
Legendary
Offline
Activity: 1666
Merit: 1057
Marketing manager - GO MP
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:31:27 PM |
|
not when it's logged. try it with numbers above and below 1.
No. Add parenthesis just to be sure. Your graph is wrong, the one at bitcoincharts is correct. oh boy ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fbitcoincharts.com%2Fcharts%2Fchart.png%3Fwidth%3D940%26m%3DmtgoxUSD%26SubmitButton%3DDraw%26r%3D%26i%3DDaily%26c%3D0%26s%3D%26e%3D%26Prev%3D%26Next%3D%26t%3D%26b%3D%26a1%3DSMA%26m1%3D7%26a2%3D%26m2%3D25%26x%3D0%26i1%3D%26i2%3D%26i3%3D%26i4%3D%26v%3D0%26cv%3D0%26ps%3D0%26l%3D1%26p%3D0%26&t=665&c=qrcVMAa4TXDsng)
|
|
|
|
oda.krell
Legendary
Offline
Activity: 1470
Merit: 1007
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:32:59 PM |
|
not when it's logged. try it with numbers above and below 1.
No. Add parenthesis just to be sure. Your graph is wrong, the one at bitcoincharts is correct. Have to agree with sarc on this one. Whether you use log n or log (n+1) for graphing, as long as you use it consistently it won't matter for the purpose this graph is used for: drawing a straight line to represent an exponential growth function. Depending on the log value you chose, that line (of best fit) will rise more or less sharply, but the actual *function* it depicts stays the same, since the line (of best fit) is based on the data plotted on the same log value. EDIT: spelling
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:36:04 PM |
|
not when it's logged. try it with numbers above and below 1.
No. Add parenthesis just to be sure. Your graph is wrong, the one at bitcoincharts is correct. oh boy ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fbitcoincharts.com%2Fcharts%2Fchart.png%3Fwidth%3D940%26m%3DmtgoxUSD%26SubmitButton%3DDraw%26r%3D%26i%3DDaily%26c%3D0%26s%3D%26e%3D%26Prev%3D%26Next%3D%26t%3D%26b%3D%26a1%3DSMA%26m1%3D7%26a2%3D%26m2%3D25%26x%3D0%26i1%3D%26i2%3D%26i3%3D%26i4%3D%26v%3D0%26cv%3D0%26ps%3D0%26l%3D1%26p%3D0%26&t=665&c=qrcVMAa4TXDsng) yeah, it's not linear. You'd have to fit a curve to that, which would predict the price would be about...ooooohhhh, say... 30ish.
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:47:46 PM |
|
To linearize things (make them 'straight' for analysis) that have points below and above 1, you add 1 to your log (i.e. log x+1).
" No you dont!
You might have accidentally done log(x+1) because log(x) +1 would not matter, it would only shift the entire graph up on the y axis while your log(x+1) totally distorts it.
You could produce even more funny meaningless curves by using log(x+42) or log(x+23) or whatever arbitrary constant you add to the price before logging. "
Actually, you could, and it still wouldn't make a difference when you reverse the transformation for prediction.
By the way, best not to keep editing posts when things have moved on.
|
|
|
|
prof7bit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:48:23 PM |
|
edit: just realised I misread you - Log (x+1) is what I did, log (x+1) was what I meant to do.
Why? It linearizes it like I said, which is one of the main reasons for logging charts.
It noes not "linearize" it, it distorts it until it fits what you want to see. Reality (this is what happened): (a) The price increased 900% from 0.1 to 1 (b) The price increased 900% from 10 to 100 if you add your $1 to all prices to distort the chart then it would seem like it (a) increased 10% from 1.1 to 11 (b) increased 900% from 11 to 101 So the gains during early period (a) are played down heavily while the gains at later period (b) remain as big as they were. This is neither a linear chart nor a log chart, its something completely meaningless (adding this $1 to all prices does not model any underlying concept or effect from reality, the $1 does not make any sense at all!) You applying dirty tricks to distort reality, to bend the curve until it matches your line, there is no reason to add the arbitrary amount of $1 to all prices before logging, you are manipulating other people with your made-up nonsense charts!
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:51:58 PM |
|
edit: just realised I misread you - Log (x+1) is what I did, log (x+1) was what I meant to do.
Why? It linearizes it like I said, which is one of the main reasons for logging charts.
It noes not "linearize" it, it distorts it until it fits what you want to see. Reality (this is what happened): (a) The price increased 900% from 0.01 to 0.1 (b) The price increased 900% from 10 to 100 if you add your 1 to distort price movements then it would seem like it (a) increased 10% from 1.01 to 1.10 (b) increased 900% from 11 to 101 So the gains during early period (a) are played down heavily while the gains at later period (b) remain as big as they were. This is neither a linear chart nor a log chart, its something completely meaningless (adding this $1 to all prices does not model any underlying concept or effect from reality, the $1 does not make any sense at all!) You applying dirty tricks to distort reality, to bend the curve until it matches your line, there is no reason to add the arbitrary amount of $1 to all prices before logging, you are manipulating other people with your made-up nonsense charts!Produce a chart to show us. then sell all your bitcoins.... edit: can I take that back, I don't want to see your chart.
|
|
|
|
prof7bit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:53:27 PM |
|
Produce a chart to show us. then sell all your bitcoins....
I have posted one. There are even more at bitcoincharts.com and many other places. All of them plot prices as they are and dont add arbitrary dollar amounts to the price before plotting them.
|
|
|
|
Richy_T
Legendary
Offline
Activity: 2534
Merit: 2243
1RichyTrEwPYjZSeAYxeiFBNnKC9UjC5k
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:55:44 PM |
|
To linearize things (make them 'straight' for analysis) that have points below and above 1, you add 1 to your log (i.e. log x+1).
No you dont! You might have accidentally done log(x+1) because log(x) +1 would not matter, it would only shift the entire graph up on the y axis while your log(x+1) totally distorts it. not when it's logged. try it with numbers above and below 1. I'm afraid I'm seeing prof7bit's side of things here.
|
|
|
|
sarc
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 98
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
June 05, 2013, 01:58:16 PM |
|
To linearize things (make them 'straight' for analysis) that have points below and above 1, you add 1 to your log (i.e. log x+1).
No you dont! You might have accidentally done log(x+1) because log(x) +1 would not matter, it would only shift the entire graph up on the y axis while your log(x+1) totally distorts it. not when it's logged. try it with numbers above and below 1. I'm afraid I'm seeing prof7bit's side of things here. Don't take my word for it, go grab your GCSE (or whatever) maths book.
|
|
|
|
|