Forums  > General  > Decreasing forum activities  
     
Page 1 of 3Goto to page: [1], 2, 3 Next
Display using:  

riskPremium


Total Posts: 16
Joined: Nov 2018
 
Posted: 2019-04-24 13:03
Hi,

As a forum newbie here I spent quite a lot of time digging through archives and couldn't help noticing there were more forum activity and more active members back in the earlier years than now. I wonder what happened.

A few hypothesis would be:
1. People don't use forums anymore. They moved elsewhere.
2. Less and less people survive in this industry as their alpha gradually decay.
3. Less and less people use this forum as they advance in their career / shift their interest.

I hope (1) is the answer but it seems (2) and (3) is closer to the truth?



longGamma


Total Posts: 13
Joined: Jan 2019
 
Posted: 2019-04-24 16:52
The industry has been in a structural decline since 08. It's intriguing to see the change over time. The user activity before the crisis with headhunters buzzing about. What a contrast. Today, crickets.

Those who remain don't have much time to spend on a forum.

Kitno


Total Posts: 392
Joined: Mar 2005
 
Posted: 2019-04-24 17:42
1. People in roles now want to keep their jobs, largely due to remuneration structural changes (high base, lower variable)
2. Regulations and Basel III has killed (through high capital costs) many areas that were evolving fast/allowed mark to model.
3. Innovation is punished (you don't get paid for it/and it's risky to do new things [see #1])
4. Fukuyama: there is nothing left to innovate!
5. Fintech was a term created to justify those who felt compelled to do something fun (trading used to be really fun).
6. Everyone moved from forums to the blockchain bro. I'm token 927848184 with MD hash: ngf7Fjfd. You can still hit me up on ICQ though: BundFlipper.

"Yeh, after that blow out I bid the bonds at 76 and you hit man...You're 77/81 now? Cool man...What? Do I care at 80? No mate... I'm 73 bid now...I'm sure you didn't just load up just for me...".

Osiris2


Total Posts: 27
Joined: Sep 2017
 
Posted: 2019-04-24 22:22
I think it's mostly a structural change in the industry. When this forum was created, there was a premium on complexity. There was a strong bid for anyone with quant expertise.
In the aftermath of 2007-8, that situation flipped and has persisted. The last decade has mostly seen complexity as a cost. Now that the legacy assets from pre-2008 are pretty much all gone, there's is not much need for the expertise that created them.
Of the people who were active here 10+ years ago, a few are still around, but most have migrated away from this forum for various reasons. The replacement rate is lower than the attrition rate because the quant finance market has been shrinking.

tabris


Total Posts: 1262
Joined: Feb 2005
 
Posted: 2019-04-25 01:29
not to mention that some folks are in positions that they cannot really be posting as much/anymore... if you look at the earlier years most of those folks are posting from their work pc as well... (myself included)

Dilbert: Why does it seem as though I am the only honest guy on earth? Dogbert: Your type tends not to reproduce.

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-26 14:21
Forum was organized back in ~2003 in response to the "no-one-can-pronounce-his-name" (W***tt). Organizers were young and energetic.

Today, young and energetic go somewhere else, but not to traditional finance. They go to blockchain/crypto, traveling and publishing, AI/ANN etc. Also, they go to W***tt. More than half of active users of NP are talking there. But that forum is not active as well.

There is no demand for quant finance from the industry. It is reduced to the high-frequency trading with building supersonic interfaces and to use of ANNs to predict the market.

Scammers also scare away users. Their presence is a demonstration of the fact that the owners of this resource simply do not care.

And as said before: Financial industry is in transition... so, the people.

New subjects are coming

For example, I am very much interested in:
- Innovations of payment industry. This landscape is changing quickly. What are business models, dynamics, pricing models, risks?
- crypto (fundamentals, pricing, risks) and
- AI/ANN (technical methods, applications). Most of answers to my AI-questions I get from stackexchange.


Is anybody busy in these? Not many.

Maggette


Total Posts: 1147
Joined: Jun 2007
 
Posted: 2019-04-26 14:42
"Most of answers to my AI-questions I get from stackexchange. "

There are smart people there, but regarding application of ML to finance, I find most of ML tribes (but not all) rather ignorant.

I use ML on a daily basis to solve problems in different engineering and industry applications. And I use deep learning and even trying to use deep reincorcement learning on a practical problem.

Many of the breath taking break throughs in ML are very important for NLP and for stuff with pictures and movies.

Finacial data represents a system that is incredible instationary and noisy. Classical deep ANN stuff suck at this. You spend so much time spending providing clues via feature engineering, that you could use a OLS regression once you have the features with almost the same effect regarding predictive power.

Back then when I started using LSTMs, GRUs and ensembles of these with other stuff the most educating thing for me was:

=> create surrogate data where you (by construction) know the underlying signal generating proccess
=> experiment how complex ANN models perform on these data sets (where you know there exists a signal)

I found it shocking how much data and complexity (depth, nodes etc) is needed for very simple patterns and how much this data hunger increases if the signal to noise ratio get's worse a little.

It is an improving field, but the above experiments pretty much healed me from "ohh...I use auto encoders that finds the representations for me in the data" idea with respect to financial data.




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

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-26 15:30

@Maggette

Most of my questions are technical, like how to use LSTM in Keras or what is the strategy to train ANN incrementally.

This man has most of it:
https://machinelearningmastery.com/develop-bidirectional-lstm-sequence-classification-python-keras/

My first NN C-lib I did 27 yrs ago for plain sigmoid-type network with back-forward propagation. No enigma. Although that ANN worked fine, nevertheless I left it behind because I was not happy with the fact that "I do not know what is happening". Today, my internal argumentation has mutated a bit -> I don't care, I need result, I need it fast and in the most flexible way...

> => create surrogate data where you (by construction) know the underlying signal generating proccess

I know underlying process too, but I also know that dynamics changes, so I have to implement kind of trigger for dynamics switch.
NN does it all... I am "a bit" surprised.

> could use a OLS regression once you have the features with almost the same effect regarding predictive power.

Totally agree. And that was the same sentiment I had 27 yrs ago, and that's why I left it behind. But still I prefer to use ANN now because most of time I spend on new search and improvement/tuning of feed/trade interface.

> I found it shocking how much data and complexity (depth, nodes etc) is needed for very simple patterns and how much this data hunger increases if the signal to noise ratio get's worse a little.

Perhaps. I noticed that there is no uniform opinion about ANN-tech. Still it works for my purpose(s).

EspressoLover


Total Posts: 381
Joined: Jan 2015
 
Posted: 2019-04-26 17:25
I'll push back and say, I don't know how much of it is finance specific, as much as it is the changing face of Internet culture.

The independent bulletin board is a dying medium. In 2003 it was the predominant mode of discussion on the web. Young people nowadays don't like curating their bookmarks to keep track of forum websites. They prefer dealing with only one or two mega networks like Reddit, Twitter or StackExchange. Just login once, then let the algorithm do most of the filtering work.

Also users like the constant dopamine hits you get from upvoting and gamification. Rather than long-lived threads with protracted discussion, the front page on these sites turns over *fast*. You get constant blue links by hitting F5. And if you agree/disagree, rather than putting in the effort to write a reply, you just click a button to feel a visceral satisfaction.

I don't really think its necessarily a bad thing that this site missed out on that transition. I could just be old and cantankerous, but the quality of Internet culture is atrocious compared to what it looked like pre-social media. I'd rather tradeoff quantity for quality. Would it be better if the forum was overrun with memes, culture war topics, recycled reposts and self-righteous pity parties? It's nice to have a little niche of the web that hasn't succumbed to the 2010s' version of Eternal September.

Good questions outrank easy answers. -Paul Samuelson

riskPremium


Total Posts: 16
Joined: Nov 2018
 
Posted: 2019-04-26 17:35
This might be a little bit off topic, but speaking of crypto i totally don't get it. 99% of the white papers I read are straightforward b**shit. Those projects are clearly scam.

As for crypto trading, it's just like forex trading but be aware of exchanges ripping you off...

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-26 18:19
@EspressoLover

Indeed, qualified personal opinion combined with occasional hang out for a beer is preferred over continuous scandals, like in Hollywood media.

All living species are destined to die. So, the forums. Amen.


@riskPremium

> 99% of the white papers I read are straightforward b**shit

So what. If crypto whitepaper b*s* is plain-vanilla:
1) it is easy to be caught, which is good
2) contribute yourself with professional stuff which other people can rely upon
3) how does it concern other 1% of people who are still delivering.

It's a market. Some win, some lose.

jslade


Total Posts: 1182
Joined: Feb 2007
 
Posted: 2019-04-26 19:59
I think forums have mostly migrated to telegram/slack/discord. This is a shame for a lot of reasons; open internet forms new communities, provides archives of valuable content and prevents shadiness. Also as others have said, foundering members have gone on to better things, or retired.

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

Jurassic


Total Posts: 253
Joined: Mar 2018
 
Posted: 2019-04-27 19:20
I dont really buy the idea that people are migrating to data science, crypto, blockchain, AI etc, as while very interesting, they arent producing jobs or industires on the same scale as finance.

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-27 22:31
In the first place, any industrial revolution is about reduction of jobs.

Jurassic


Total Posts: 253
Joined: Mar 2018
 
Posted: 2019-04-27 22:56
@nikol good point but reduction is in finance not blockchain, crypto etc

Kitno


Total Posts: 392
Joined: Mar 2005
 
Posted: 2019-04-27 23:19
A tad ironic that trading is dead from regulations banning leverage (and frankly if you look at the pre-IPO days of before 1998 it's just turning the clock back to how trading was in the partnership merchant banks, even if the bond market was smaller - and so as a system not analogous to today).

Anyway, you can still make a ton of cash in banking but it's in M&A and corporate finance, which is crap.

"Yeh, after that blow out I bid the bonds at 76 and you hit man...You're 77/81 now? Cool man...What? Do I care at 80? No mate... I'm 73 bid now...I'm sure you didn't just load up just for me...".

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-27 23:40
Crypto, blockchain, AI help to make processes in finance more efficient, hence finance (quant) jobs shrink. Quants out of jobs naturally flock into crypto, blockchain, AI by accelerating the revolution. Reduced non-quants (backoffice, bank retail officers, compliance officers etc.) I don't know where they go...

Couple of years ago I saw Google advertising free on-line marketing courses. I concluded that they see an opportunity to boost this kind of activity and hence the on-line markets with all those cloud based AI processing power, Big Data Analytics...

World Economic Forum warns - "Jobs in 20s are freelance"

Huobi has made toolkit to franchise Crypto Marketplaces. Take it and startup. Nowadays I try to help people willing to disrupt payment landscape even beyond what does exist today (like XRP, those Bank coins, Kantox, Revolut etc.) Not sure, if these people succeed.

So many things are happening!

Jurassic


Total Posts: 253
Joined: Mar 2018
 
Posted: 2019-04-27 23:44
@nikol your ideas are really for startups not for getting a job (which is what a quant was)

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-28 00:10
They call it now "Data Scientist"

Kitno


Total Posts: 392
Joined: Mar 2005
 
Posted: 2019-04-28 00:18
Also, what is a lot of money?

I was told at 21 I'd be on a bar a year by 30 (obviously I was in a shit bank).

If you are solely driven by money then you should be an entrepreneur (the French should get a word for that). If you yearn for sovereignty and upside well trading used to offer that, now it doesn't.

If you can have an easy job doing 8-6 earning 250K in your mid-20s you really should be happy; it's a pity the CBs made money free or you'd be able to buy 4-beds in Kensington like in the early 90s.

Life is odd. I just got a guy I worked with for 18 months 9 years ago a job yesterday.

"Yeh, after that blow out I bid the bonds at 76 and you hit man...You're 77/81 now? Cool man...What? Do I care at 80? No mate... I'm 73 bid now...I'm sure you didn't just load up just for me...".

Jurassic


Total Posts: 253
Joined: Mar 2018
 
Posted: 2019-04-28 10:31
@kitno how many people get the entrepreneur option to work and get paid more than in a bank?

Kitno


Total Posts: 392
Joined: Mar 2005
 
Posted: 2019-04-28 15:45
@Jurassic - I didn't say they'd be successful at it!Wink

"Yeh, after that blow out I bid the bonds at 76 and you hit man...You're 77/81 now? Cool man...What? Do I care at 80? No mate... I'm 73 bid now...I'm sure you didn't just load up just for me...".

nikol


Total Posts: 770
Joined: Jun 2005
 
Posted: 2019-04-28 16:12
@Jurassic and @Kitno

It is not about success rate over the large sample, but about how do you youself place within the sample.

Nonius
Founding Member
Nonius Unbound
Total Posts: 12786
Joined: Mar 2004
 
Posted: 2019-05-02 11:21
was chatting with one member on a "social media" venue. to add to EspressoLover's point, this other member put it brilliantly (comparing the old anonymous message boards vs social media).

"With anonymity: post content was the only way to develop credibility. The message was everything. No one pre-judged the messenger. And only by getting immersed in others’ posts, we evaluated who was credible or not.

Now: profiles are used in an attempt to create credibility rather than content. It’s a pity. But perhaps another example of fashion over form.

I miss the old days too. But maybe that’s just more a sign I’m getting old!!!"

Chiral is Tyler Durden

Osiris2


Total Posts: 27
Joined: Sep 2017
 
Posted: 2019-05-02 22:45
NP needs a twitter handle and an insta feed
Previous Thread :: Next Thread 
Page 1 of 3Goto to page: [1], 2, 3 Next