Pages:
Author

Topic: : - page 10. (Read 70881 times)

hero member
Activity: 616
Merit: 500
Firstbits.com/1fg4i :)
October 14, 2011, 04:17:25 PM
I see, alright.
hero member
Activity: 616
Merit: 500
Firstbits.com/1fg4i :)
October 14, 2011, 03:16:27 PM
I wouldn't mind keeping things more informal; i wasn't planning on spending an amount big enough to make it feel like serious busyness anyway (just one or two BTC), and i wouldn't really hold it against you if you took longer than i expected to return the money. But if you still don't wanna take on the responsibility, even with little amounts and without a hard deadline, i understand; i too tend to feel somewhat uneasy with putting myself in the position of having others counting on me for important stuff.
legendary
Activity: 1666
Merit: 1057
Marketing manager - GO MP
October 14, 2011, 01:17:44 PM
Here is a video of an (awesomely pretty  Shocked) Professor Elisa F. Long of the Yale School of Management explaining Bayesian Analysis...
http://www.youtube.com/watch?v=bX6oWT707UI
sr. member
Activity: 462
Merit: 250
October 14, 2011, 12:32:23 PM
Ha ha.  Very humble of you. Smiley
sr. member
Activity: 462
Merit: 250
October 14, 2011, 05:34:49 AM
The tools described in a book like BDA actually give you more flexibility to innovate.
hero member
Activity: 784
Merit: 1000
bitcoin hundred-aire
October 14, 2011, 12:12:27 AM
Can i send you a little bit of money and you add it to your investment process, and after you made more money you return it plus a fraction of the profits proportional to the amount i contributed, taking away a percentage of the profits to yourself if you want of course?

I'd also do this if I could see how much profit this has made you, chodpaba.
sr. member
Activity: 462
Merit: 250
October 14, 2011, 12:11:05 AM
I do Bayesian data analysis for a living.  The bible is Gelman et al.'s Bayesian Data Analysis.  Assumes a solid grounding in probabilistic reasoning.
hero member
Activity: 616
Merit: 500
Firstbits.com/1fg4i :)
October 14, 2011, 12:09:37 AM
Can i send you a little bit of money and you add it to your investment process, and after you made more money you return it plus a fraction of the profits proportional to the amount i contributed, taking away a percentage of the profits to yourself if you want of course?
hero member
Activity: 784
Merit: 1000
bitcoin hundred-aire
October 13, 2011, 11:09:39 PM
Bears make money, bulls make money, but pigs never make money.  I keep having to re-learn this lesson.  Cry

That said, those probability charts are *STILL* eerily accurate for getting the likely direction of the day's trading.

That is because it is based simply on how the market has behaved given similar circumstances. This should be surprising in no way. People have been trading successfully on Bayesian analysis since the 1940's, still very few seem to understand it. The particular weakness of Bayesian analysis that I have struggled with most in this market is simply the lack of history. The more history you have the better it works. And we are just on the cusp of having enough history to make it very, very effective. I may get to a point before too long where I may have to stop publishing the output.

Is this methodology outlined in a clear, concise summary with pointers to further reading somewhere?  I'd like to read about it... might make a bot lol.
member
Activity: 99
Merit: 10
October 12, 2011, 08:42:01 PM
Thursday.

Looks like a snapshot before the spike to 4.45.
legendary
Activity: 1666
Merit: 1057
Marketing manager - GO MP
October 12, 2011, 04:04:08 PM
Bears make money, bulls make money, but pigs never make money.  I keep having to re-learn this lesson.  Cry

That said, those probability charts are *STILL* eerily accurate for getting the likely direction of the day's trading.

That is because it is based simply on how the market has behaved given similar circumstances. This should be surprising in no way. People have been trading successfully on Bayesian analysis since the 1940's, still very few seem to understand it. The particular weakness of Bayesian analysis that I have struggled with most in this market is simply the lack of history. The more history you have the better it works. And we are just on the cusp of having enough history to make it very, very effective. I may get to a point before too long where I may have to stop publishing the output.

Interesting, but very academic. Difficult to understand for the layman. (At least what I find on wikipedia)
Would you share some read to better understand the concept?
legendary
Activity: 1666
Merit: 1057
Marketing manager - GO MP
October 12, 2011, 03:49:42 PM
The performance of your analysis was surprisingly disappointing the last few days. Maybe I am just too greedy but none of my orders were filled...

The big boys are obviously watching this and make sure it doesn't happen  Undecided

The analysis isn't the half of it though.

What really makes the difference is how aggressively or conservatively you trade on the data. This is a very individual thing depending on your positions. I do not have a single position, I have longer term and shorter term positions. The longer term positions have been mostly standing pat, but the portion I have allocated to shorter term positions allow me to trade profitably on a nearly daily basis, on essentially the same data, since it does not change all that much. As well, I would remind you that the analysis is tuned to TradeHill data, it will not necessarily provide the best targets for Gox.

My guess is that you are probably trading form a single position that is keeping you priced out of the market at the most probable trades. You might want to review how you are doing money management and diversify a bit.


Your suspicion is correct, I am on mtgox with a single position.

The thing is though that I learned that trading with such a small amount in more positions isn't that rewarding. I would have made 25% with a single position when I started but have made 2 positions one were I would have gained 25% and one with 27%. The first was assuming a idealized buy the latter where I actually bought at. So I changed my strategy but was always outpriced..

I think I will return to my own strategy again, since if I would have stayed at it I would have made some profit during the time...
full member
Activity: 154
Merit: 100
October 12, 2011, 03:39:58 PM
Bears make money, bulls make money, but pigs never make money.  I keep having to re-learn this lesson.  Cry

That said, those probability charts are *STILL* eerily accurate for getting the likely direction of the day's trading.
legendary
Activity: 1666
Merit: 1057
Marketing manager - GO MP
October 12, 2011, 03:05:23 PM
The performance of your analysis was surprisingly disappointing the last few days. Maybe I am just too greedy but none of my orders were filled...

The big boys are obviously watching this and make sure it doesn't happen  Undecided
legendary
Activity: 1246
Merit: 1076
October 11, 2011, 09:35:30 PM
Wednesday.

*image*
Is this painting a bleak picture? Unless the 2% happens, it looks like there won't be a rally...
legendary
Activity: 1386
Merit: 1000
October 11, 2011, 01:35:53 AM
Tuesday

How to read this graphics? Looks like today price should fall to 3.80 ?
legendary
Activity: 1246
Merit: 1076
October 10, 2011, 10:55:42 AM
Here the high price median got squished together with the lower quartile at $4.19.

Monday.
*chart*
Is the model probability calculated directly off support/resistance? If so, then is the overlap due to intense resistance at $4.19?
legendary
Activity: 1666
Merit: 1057
Marketing manager - GO MP
October 09, 2011, 02:39:41 PM
Just wondering where is the average high price probability in this one?
The blue thin line seems to be missing is it at 4 or 4.15?
sr. member
Activity: 462
Merit: 250
October 08, 2011, 07:14:05 PM
replying so I can find this thread easier Tongue
legendary
Activity: 1246
Merit: 1076
October 08, 2011, 06:32:48 PM
No, but at these levels, relevant pricing data is getting pretty old.
Even a couple years into the future, $30 would still be a preety powerful resistance (unless the price goes back over $30).

Edit: fixed quotes
Pages:
Jump to: