Forums  > Trading  > Quantopian – why I don’t take part  
     
Page 1 of 1
Display using:  

finanzmaster


Total Posts: 118
Joined: Feb 2011
 
Posted: 2017-01-12 23:38
https://letyourmoneygrow.com/2017/01/12/quantopian-i-dont-take-part/

Quantopian is a very interesting FinTech project for virtually everybody, who wants to try the algorithmic trading. Yet I explain why I myself - a successful trader, experienced quant and good programmer - don't take part (and what Quantopian should undertake to make me do).

NP fellows, especially those who do take part, are very welcome to discussion!

www.yetanotherquant.com - Knowledge rather than Hope: A Book for Retail Investors and Mathematical Finance Students

a路径积分


Total Posts: 77
Joined: Dec 2014
 
Posted: 2017-01-13 04:55
I have no right to decide what community guidelines are on this and others seem to be kinder towards your posts so I've steered clear of speaking my mind.

However, you asked for feedback, so I'll be honest.

This is probably the umpteenth time you've posted a link to your website with the equivalent of Facebook clickbait. It would be nice if you could actually share your thoughts on here rather than use the forum to monetize your website. I respect that you are very passionate about your views and appreciate the content you're trying to create, but it creates unnecessary noise and distance between us if I have to click your website URL each time to engage in a discourse with you.

Regarding your topic - many of us here have amazing faculties or technical abilities and can do the work ourselves. Others depend more heavily on marketing, hard work, and crowdsourcing. Quantopian is not my kind of thing either and the shortcomings are obvious to us, but I don't feel a need to pick up a loudspeaker and put them down.

pj


Total Posts: 3317
Joined: Jun 2004
 
Posted: 2017-01-13 07:09
Well, since the site is solely finanzmaster's and all
the "monetization" there is to see his logo
I do not see any harm in that.

Or should he post only here (to a restricted community)
or only there, or post twice?

Moreover his posts there are quite interesting.
And you are welcome to discuss them here.

So what is your problem?
The extra click?

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

jslade


Total Posts: 1064
Joined: Feb 2007
 
Posted: 2017-01-13 07:46
I thought it was worth the click for the new Russian saying to add to my collection.

Crowdsourcing anything is kind of stupid. Someone just asked me to join the Kaggle contest to find lung cancer in MRI images, "for fun." OK, I could probably make a good effort at that. But if I wanted to work on that kind of problem, should I spill my guts out and give away my code to Kaggle (who will probably award the win to someone who data mined their loss function), or ask for an angel investment where I get to keep real profits and build a business that means something?

"Learning, n. The kind of ignorance distinguishing the studious."

EspressoLover


Total Posts: 221
Joined: Jan 2015
 
Posted: 2017-01-13 08:02
> should I spill my guts out and give away my code to Kaggle (who will probably award the win to someone who data mined their loss function), or ask for an angel investment where I get to keep real profits and build a business that means something?

Never underestimate the extreme lengths that modern-day 25-year old geeks will go to avoid making a phone call or putting on a suit.

Maggette


Total Posts: 941
Joined: Jun 2007
 
Posted: 2017-01-13 11:40
I agree with PJ on the "click the link" point.

On Quantopian, Kaggle etc: my partner and myself run a small company that provides services and consulting in the area of machine learning/predictive analytics, operations research/optimization, scientific computing and classical software development.

We do think about using these competitions to show off a little and as a marketing tool (content for blog posts, our Github repo).

Sure. The stuff I have seen so far on kaggle is IMHO is only loosley related to the predictive analytics projects I did in the "real world". IMHO you need a different skill set there. And I think most professionals in that area know that and aren't to impressed on stuff you did on kaggle. A little bit like certifications in software products.

On the other hand. I do have a lot of blind spots when it comes to machine learning and statistics. I have a good theoretical background and lots of practical experience in some areas, but lack theory and hands on experience in other fields. But iI want to develop in these fields.
For example: jslade posted stuff about conformal prediction and compression a while ago. Or NLP. I think you can use some competition data set to get your feet wet on a certain topic. I would spent time on it anyways.

The hard thing is to carve out the extra time. I work 50 to 60 hours a week,have a girl friend and want to work out 3-5 times a week. And we do want to contribute on open source software projects too. So time most definetly will be the major constraint here.

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

mtsm


Total Posts: 193
Joined: Dec 2010
 
Posted: 2017-01-13 16:29
finanzmasters blog posts on ML, DL as well as on Quantopian are a bit naive and a bit half baked. No offense, but it shows that he is not in the mainstream or the industry.

As far as Quantopian goes, I read your post, and every bit you quote is standard hedge fund talk. You can hardly expect anything else. You may not agree with it, but you won't find anyone telling you to join the fun(d) with a strategy that has high beta to market, a high correlation to peer strategies, a volatile PnL, low Sharpe ratios, a 50% incentive payout and other nonsense...

finanzmaster


Total Posts: 118
Joined: Feb 2011
 
Posted: 2017-01-13 19:33
@pj, @Maggette
Thank you very much for your moral support.
I don't feel that I behave badly on NP but in either case it is important for me that at least some NP-patricians welcome my activity.

@a路径积分
Believe me if I wanted just (quick) money I would have moonlighted writing a code or giving tutorials. You will never get rich from google-ads and amazon affiliation on such a niche website like letYourMoneyGrow.com

@jslade
Cool that you like Russian sayings. Yes, Russian is very expressive.
This saying about the secretary can be even complemented: "And, additionally, she should work completely naked" (I mean, Quantopian requires a complete disclosure of the algos).

@mtsm
>No offense
Not a milligram :)

>but it shows that he is not in the mainstream or the industry.
It depends on what you mean. If you mean that I never worked for a hedge fund, this is just a fact, I never did (Maggette, are there hedges funds in Germany at all?).

But if you mean my skills, than let me say another Russian saying:
"dear sir, you deigned to being wrong".
You know the performance of my fund, don't you? https://goo.gl/GZJXW3

Though we have few (if at all) hedge funds in Germany, we do have enough mutual funds (mostly with a miserable performance).
The chart below shows the performance of the best funds of DEKA and Union Investment (two big German financial groups) and Dirk Müller & Max Otte fonds (self-named stock market gurus). As you see, none of them beats the DAX in 2016 (let alone my fund). And also in past years nearly all of them lost to the DAX.


BTW, there is a very interesting discussion on Quantopian, some of my cautions are (partially) retorted.

www.yetanotherquant.com - Knowledge rather than Hope: A Book for Retail Investors and Mathematical Finance Students

goldorak


Total Posts: 979
Joined: Nov 2004
 
Posted: 2017-01-13 20:30
@espressoLover: > Never underestimate the extreme lengths that modern-day 25-year old geeks will go to avoid making a phone call or putting on a suit.

I would nominate this as "quote of the year" for 2017. Applause


If you are not living on the edge you are taking up too much space.

finanzmaster


Total Posts: 118
Joined: Feb 2011
 
Posted: 2017-01-13 20:33
[offtopic]
Maggette> want to work out 3-5 times a week

Ich bin neidisch - I am envious!

As I was a student I was able to lift my own weight in the upright position and now I just go a couple km on foot to office and occasionally make 10 pull-up.

Anyway, I would say a workout 3 times a week is enough. Otherwise your body will have no time to recover. As a student I did pumpen-pumpen-Schwarzenegger 3 times a week, 2 hours, two muscle groups per training.
[/offtopic]


www.yetanotherquant.com - Knowledge rather than Hope: A Book for Retail Investors and Mathematical Finance Students

Maggette


Total Posts: 941
Joined: Jun 2007
 
Posted: 2017-01-13 23:35
"(Maggette, are there hedges funds in Germany at all?)."
Yes, there are a couple hedge funds and prop shops

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

finanzmaster


Total Posts: 118
Joined: Feb 2011
 
Posted: 2017-01-14 10:24
Apropos Prop-Trading, there is SSW-Trading GmbH near Hamburg (glorified also for their job-ads that remain the same years long)

Since SSW-Trading is a PTC, little known about them but thanks to bundersanzeiger.de one may get some info.
There is no info on their trading performance in the canonical sense but at least we can estimate the ROCE. And we readily see that it fluctuates heavily (though there is no obvious dependence on market). In this sense even is a strategy is more or less market neutral, it doesn't yet imply a consistent performance. Moreover, the performance fluctuation due to whatever factors can be larger than teh performance fluctuation due to the beta.



www.yetanotherquant.com - Knowledge rather than Hope: A Book for Retail Investors and Mathematical Finance Students
Previous Thread :: Next Thread 
Page 1 of 1