Wall Street still likes Quants

I read this interesting article in Bloomberg: Link Here

Engineers and Quants have gotten a bad rap along with their backtesting and statistical techniques recently.

Take heart. Wall Street still loves Quants and their computer power continues to grow.

I don’t know what they are plugging into their computers. But that is okay because the don’t want to tell me. It may not be just classical financial theory however.

Edit: sorry about the initially wrong Bloomberg link

Actually, from my sense of the article, it seems like just a more souped-up version of what we do right here on p123. We make quick and efficient use of structured data that for most of Wall Street history was inaccessible and/or cumbersome and hence unused by most investors. Technology changed and we’re tapping into that. The hedge funds described in the article are doing the exact same thing just with different types of data and different platforms.

The Bloomberg article focused on a very narrow segment of this overall trend, but actually, we’re all right smack in the middle of it. And as is the case with us, there will be differences in approaches, opinions, and outcomes there too. And the newer unstructured-data trends discussed there are, indeed, much newer and it will likely take a while for users to figure out how to make it succeed. Note that so far, all the press it has gotten had talked about the golly, gee whiz aspect of the technology. We’re not hearing about the alphas the users are actually earning.

Note, too, that the main news sources that publish such articles, Bloomberg and Reuters, are the same organizations that are trying to make money by selling access to this sort of data. Coincidence? Hmmm. :slight_smile:


That is kind of my point. Not all of us, but some, are using backtesting and sims in an appropriate manner. Not me, I’m a quant wannabe.

We do not need to make our computers dumber with added noise if we want to continue to be in the lead. Wall Street won’t do that.

It is probably me that worries about a 1 to 2% difference in sim backtests. That is wrong. I promise to get back on my medications when I start using decimal points in my calculations.

Keep up the good work!

Edit: I have also read about Shaw and Ed Thorp (Thorp since I could read). Private, highly successful and we still know little about Shaw. I have read that some things that used to work no longer work because Shaw is exploiting the market inefficiencies on a large scale.


Agree that P123 let’s normal retail investors play in this space.

For what it’s worth, I have seen many internal traders at some of the big quant shops with 5-7 year live trading records and realized out-of-sample 2 Sharpe ratio’s and 20% plus annual returns (in some cases over 40% AR for more than 7 years - documented by people I know who’ve done due diligence). Many of them go out on their own as prop traders. You don’t hear about them in the popular press, because they don’t advertise much or have web-sites (often) and they aren’t seeking or accepting retail investors. They are funded by large seed funders and super HNW’s mostly. I can’t get into most of them due to the min’s (often $5MM or $10MM or more), but they are out there and I’ve seen them and know people that invest - and many of them do amazingly well.

Many of them are trading short-term and very short-term systems. Many are trading at least a large number of systems that have nothing to do with fundamentals. At the 1 and 5 min levels, how can they? Many of them are using the same methods you dislike so much, or seem to.

Apart from exceptional math and programming skills / teams, one key to what they are doing is having data that is a) really, really high-quality (I met a team of 2 guys from MIT that spent 3 years and a lot of money cleaning and preparing their data set before they even started) and b) data that is not everywhere. They are often paying for, or generating, really unique data sets. They are often using real-time analysis of order book dynamics and changes in those dynamics (bids vs. asks), for example, or proprietay earnings upgrades.

I don’t think P123 users can compete with them. They are pretty amazing and very sophisticated and coming out of the best prop trading shops where they’ve built and honed these models with millions to billions of dollars of tools / data behind them.

One interesting thing for P123 to explore is, in addition to R2G, have a ‘data license’ area where people can post proprietary data for sale to model builders - with very strong documentation on why and how it’s PIT and where it comes from. It may turn out to be useless, but it may allow P123 users to ‘import’ or run systems on whole new data sets. These would be additional monthly license fees for users - so it’s an additional revenue stream for P123 - and can bring in data vendors. A key would be really documenting why and how the data works.

A core issue with what P123 is trying to do now is it’s trying to be a ‘one stop shop.’ I believe R2G is better as a stand-alone business.




Your posts are always insightful. In the examples you mentioned, can you comment on the liquidity of the stocks these guys trade. Are they super low liquidity like the popular R2G models here? Or are the short term models much higher liquidity? thanks.


I only see the people taking outside money and they are not trading low R2G level liquidity, but are low’ish.

On the stock side, the ones I have seen this year (2 of them), one trades 100’s of stocks (500 or so I think), the other trades 2500 stocks on average (US, they trade 5000 if you want them to trade Europe and Asia). They both have about $50MM in assets. Their returns are reported by third party administrators and they are audited as well. One was trading medium hold fundamental systems (3 month average hold), the other was trading 5 day average hold. They both claim they have fairly large capacity (up to a billion dollars) across their ‘portfolio’ - but the port is many systems together.

The one with 2500 stocks has machines building systems in specific silos - mean revert, momentum and fundamentals - and then an algo/system that selects and weights from among the subsystems based on their projected forward performance and correlations, and then another set of systems for adding dynamic hedge overlays. They are not just ‘hoping’ their models (or data) will work. They have a 1.9 Sharpe from 2004 to 2010 (when they started on their own). 10 Phd’s on the team. Over 3000 models - each of which is ‘simple’ on it’s own - very few factors.

But, I only see the one’s who are taking outside capital. I know a Phd Finance Professor in Quant Finance who runs a quant trading program of his own. He’s also met many people and reviewed their live numbers. He tells me that some of these people have put up 40% to 50% returns for more than a decade with their own money. I have no idea what liquidity they are trading.

But, as for me, the very high return people I’ve seen are mostly in futures - and tend to blow up over time, unless they are under 20% AR. But I have seen 10% to 20% AR from prop traders going out on their own with 2 Sharpe’s. Although, it’s very rare (Sharpes in .7 to 1 are much more common for the best traders). I would invest with these people in a second if I could. I am not in their league - at least not with the tools I have and the time I’ve spent so far. But, it’s very irritating to hear that simple technical systems are not real quant. They are. They are not the only way, but they are a best practice for a very elite group of traders. Again, I’m not in their league in this regard, but they are world class.


Very interesting. Thanks.


What’s this . . . another spitball tossed at the idea of developing strategies based on well established common sense notions of how stocks are priced? Do you actually know what I’m talking about? I wonder because frankly, the vocabulary you use is along the lines of what I would expect from one who really does not understand what this is about. I find it very difficult to respond to your posts (as with the shots Steve takes along similar lines) without being rude (I really try and hope I don’t slip, at least not too often and I regret when I do) but mindful of the fact that there are more members here, many new, who have access to and see these forums than the number who post, and the dangers of allowing what amounts to nonsense to stand without opposition, especially nonsense that cold hurt users, could cost them money.

We know we have to enhance our forum/blog capability and in an ideal world, I’d be tons of content on p123. But as you know, we have some big projects that are very important to many members on the first burner. So for now, I’m writing elsehwere. a couple of days ago, I posted some links to TalkMarket.com where I posted some articles in a series that discusses why stocks are prices are what they are. By now, there’s a third, and I’ll send the fourth and final article after I get a chance to self edit. Can I suggest you (and anyone else who is interested in this topic) read them and then come back here and take up the discussion?

For quick reference, here are the first three urls:

Part 1

Part 2

Part 3

Part 4
Forthcoming soon

I vote that we let Tom use whatever dumb ideas he wants as long as he does not complain about minor changes in his sim results too much (which he hasn’t).

I vote the same way . . . and won’t even include the proviso that he not complain . . . .

People are free to use any dumb ideas they want, and even advocate for them on the forums. But surely you are not voting against the offer of an opportunity to convert dumb ideas into smart ideas. :slight_smile:

Marc - there is s reason I am rude. A person with your position within P123 should be fair to all, keep an open mind, and not run rough shod over anything that doesn’t fit into your personal agenda.

This is a view written by fundamentalists that have a much less biased view than you:


Now quantitative analysis has been around for a long time and most large financial firms have q.a. departments. It is not your position to decide everything is garbage except for your opinions. It is your job to support P123 members in whatever fashion they choose to use the tools on the site. There are many problems with this site, bugs that aren’t being fixed, people are screaming to have basic tools improved and issues addressed. I know of people entering 300 trades every Monday because of issues with the book technology. People were promised European markets how long ago? Who is supporting Newbie questions on the site? Why are you not addressing real issues instead of forcing your beliefs on everyone else?

Now, you say that I don’t have an open mind. Well that is not true. I believe you have a lot to offer in the way of fundamental analysis. But my mind closes when you run down everything in sight except for your religion. All I am hearing is an agenda which will likely relegate P123 back to stock screener technology with one purpose in mind. Valuation techniques come and go. If you don’t believe me then go to AAII and see all 70-something screens that were once the best thing ever, most of which are underperforming.

Take care

We can definitely let Steve do what he wants to. I would not be able to stop him anyway. As far as him learning anything: hopeless! LOL

Love you Steve and Marc too.


This is a quote from your articles I do agree with:
“Disdaining someone else for the style they choose makes as much sense as disdaining one who prefers Merlot to Chardonnay.”

The founder of XXXX Capital achieved a near 2 Sharpe over 10 years while working at Millenium and Worldquant and managing a staff of 40. He did it with statistical arb. strats with 20% daily portfolio turnover. Are you saying that you understand quantitative investing better then him or that his system is flawed and worse than yours?

I don’t think so, I think you are saying that most non-math geniuses shouldn’t try to copy him (it is a him), and I agree with that. But it sure seems like you are calling the above trader dumb?

And yes, I do know how to value a business. I built and sold a fairly large company. We prepared multiple valuation models for our business. We sought buyers. The buyers then prepared their own competing valuation models of the business based on how they could integrate our business with theirs and what that might be worth to them - and their desire to make money and told us all our valuations were wrong and this is what they would pay. We then negotiated and eventually reached a sale price. But it took years - because we both had to really want to do the deal. There was nothing in their analysis that considered the future ‘dividends’ they would receive. They were looking at the total returns they thought they could earn cash on cash.

I have also looked at or been pitched by several hundred start-ups and been on the board of a half dozen or so angel or venture tech backed companies. So, I think I know how to grow a business fairly well - and understand the basics of valuation. I also understand fairly well how businesses run and what works generally and what doesn’t. What matters most in VC deals is the size of the opp and quality of the team, AND the amount of capital chasing the deals. The macro ‘deal cycle’ is the single greatest driver of returns in this asset class, none of these relate to the future dividends.

So, I don’t agree at all with this statement or the many pages based on it.
“There is one destination at which all should aim, the present value of future expected dividends.”

Stock is ownership in a company. It is also a giant multiplayer game and recognition of the rules of that game.

At a bare minimum, I would strongly consider buying a stock (or company) I thought has $1000 million in assets and no earnings or debt or future ability at earnings or dividends, if it was selling for $100MM. This has nothing to do with earnings or dividends. It would have to do with the fact that I’d expect someone else to eventually pay something way more than $100 million to buy it’s assets, and I would expect to be able to make the difference.

I might also consider buying a stock that is surging upwards and has strong supporting volume surge and increasing positive tweets or news mentions, if I am a short-term momentum trader with a short holding period. I would know that this has nothing to do with future expected dividends. It’s based on statistical probabilities of other market players pushing this type of stocks price higher.

If I’m holding any public stock, all I really care about is the expected total return of the stock relative to the ‘volatility’ and downside risk over my anticipated holding period. In most cases with most current public stocks, that is more reflected by the predicted future market price changes of the shares, not expected future dividends. But in all cases, it’s the sum of dividends and price changes in the stock. To leave out a detailed analysis of possible causes of the expected future ‘price changes’ for an investor is a huge gap in your article.

And yes, I have read finance text books and thousands of academic articles and am aware of the foundations of what you are drawing on.

As another example, if I am a fundamental investor and think a company has a huge amount of strategic value (i.e. a customer database or ‘blocking value’) and/or assets a company might want to aquire - and that value is much more then the current stock price, and I think there are 2 or 3 potential companies that will start a bidding war and bid up it’s price, I will buy that stock - even if I expect future dividends paid out to be zero and earnings to be negative. This happens all the time - either for competitive blocking reasons, or talent grab reasons or IP gains, etc. People buy intellectual capital or teams or other assets. If I am an M&A guy, I may very well take speculative positions in these types of companies.

Good luck with your articles.


I read Tom’s entire thread with great interest and his comment’s were meant to be informative. He came here to share his knowledge and experience with this community. Your comments, on the other hand, came with a warning! As well they should. If sharing ideas will be greeted with distain, I think you’ll find fewer members contributing to these forums.

As I stressed last week in a different thread, these comments are viewable by the public per your stated present and future plans for P123. My original thought was that RTG designers sould expect that the things they ‘say’ on this forum will be effectivly: ‘on the record’. And that new members subscribing to their RGT portfolios will take their forum post comments and demenor into account before subscribing to their ports.

Bobby Jones after watching Nicklaus win the 1965 Masters: “Nicklaus played a game with which I am not familiar.”

This is the essence of how I feel about the article. I’ve seen glimpses into the world of sustained Sharpe 2+ strategies on size. I have a great admiration for TGS Management, Two Sigma, Renaissance, Hutchin Hill, DE Shaw, Winton, Waterfront International, Prediction Company… They are pros with resources to be admired as one would admire a great golfer. We’re both hitting a ball with club, but I humbly feel that the difference between what they do and what I do is the difference between finishing top of leaderboard in the Masters and mini-putt.

Marc and Tom,

I have thoroughly enjoyed the generosity of knowledge shared by both of you and have learned a great deal. In fact, I have learned more about investing on this forum (from the usual suspects) compared to getting a Masters in Finance!

Please keep the faucet of knowledge and experience flowing and don’t let personal agendas get in the way.


Thank you to all of you (Marc, Steve and Tom) for your good posts and sharing your insights.

I personally believe there are many ways to become successful in capital markets. Both statistical and fundamental. In fact, I think the best portfolios are a combination of statistical and fundamental strategies. Typically fundamental strategies have a short volatility profile (means they underperform short-term when volatility increases, because investors focus on liquidity and technicals and not on longer-term fundamentals), while statistical strategies typically have a long volatility profile (means they outperform short-term when volatility increases for example because of wider spreads).

I got to come across many great multi-strategy firms and FoHFs that combine great market neutral strategies across neutral, long and short volatility profile strategies. Exceptional firms in this space are Millennium, Visium Global and Balyasni. But also DE Shaw, and most other really large HFs combine multiple uncorrelated strategies.

Focusing on fundamental strategies and single time frames only will lead to worse Sharpe Ratios and larger drawdown because this approach is just not robust enough across market cycles.



  I have made a living for 30+ years from the financial markets.  I am still totally amazed at the variety of ways people make money in them.  The value of the forums is only enriched by the different viewpoints.


Agreed! Beware of monocultures! Forums, like the markets, operate best where there is a diversity of opinion. Personally, I prefer to combine both fundamental and technical analysis. I just wish P123’s technical tools were stronger since it is the best place I have for testing new ideas. And while I consider myself open-minded, I recognize that I carry biases, too. I just read an article - http://finance.yahoo.com/news/10-hedge-fund-supercomputer-thats-090001837.html - that talks about machine-learning efforts on Wallstreet. Peew!


I presume you did read the posts, right? (I just submitted #4 and based on some other things you said, I think you’ll definitely want to check that when it’s available.)

Let’s not be so quick to dismiss. Let’s explore this topic, which is actually a broad and fascinating one.

Then as you would know, the idea of a stock being worth the present value of future dividends was not presented as a specific formula anyone would actually use. (I can’t recall if it was Part 1 or Part 2, but I even presented a framework for adapting this to non-dividend payers.) Essentially, it’s a summary of the overall idea that you won’t invest in anything unless you expect a positive return (with the extent of positive returns being a separate issue that was briefly introduced but not explored because it’s too big a topic for that particular series). It’s a framework, not an instructional manual. For example . . .
[/quote]"There is one destination at which all should aim, the present value of future expected dividends."nother example, if I am a fundamental investor and think a company has a huge amount of strategic value (i.e. a customer database or ‘blocking value’) and/or assets a company might want to aquire - and that value is much more then the current stock price, and I think there are 2 or 3 potential companies that will start a bidding war and bid up it’s price, I will buy that stock - even if I expect future dividends paid out to be zero and earnings to be negative. This happens all the time - either for competitive blocking reasons, or talent grab reasons or IP gains, etc. People buy intellectual capital or teams or other assets. If I am an M&A guy, I may very well take speculative positions in these types of companies.

Suppose you do want to buy a customer database, for, as you suggest, “blocking value?”

Aren’t you then expecting a positive return on the asset. It won’t be as simple as expecting the database itself to generate a cash return to you. But you would not spend the omen unless you expected some sort of positive return. Presumably, by mentioning blocking value, you expect your ownership of the database to frustrate competitors and lead to a larger market share for you. That incremental market share would, obviously, translate to additional revenues and presumable on to additional profit. And let’s assume you borrow money to finance your purchase of the database. Wouldn’t you want the additional profits you gain through higher market share to exceed the cost of the debt you take on to finance the asset.

That is exactly what the present-value of dividends framework is about; linking your investment and the price you pay to an economic return.

As the series explained, the links are not always simple and direct. And they aren’t always spelled out in specific terms by those making the decisions. For example, in your business, you may have accumulated enough experience to recognize what the database can accomplish and whether you’d be getting good deal or overpaying for it even without spelling out the details of the framework and calculating step by step. But what you do consciously and what you do implicitly all boils down to one thing; the purchase will be good if the present value of the economic return you get (whatever the case-specific details and however precisely or imprecisely you measure it) exceeds the economic cost. \

So if you hire an MBA to evaluate the purchase and recommend to you whether you go or pass based on a particular asking price, wouldn’t you expect that the case he makes to you will somehow brother talk about issues relevant this; the cost of money used to make the purchase, what impact it will have on your business, etc.? Would’t you hope that all the homework he does somehow or other aims at those questions?

I don’t have experience buying and selling businesses privately, but as a former junk bond manager, I do have experience financing such purchases so I do know the issues creditors expect to addressed. And I’ve seen and, sadly, experienced the ugliness that happens among those that don’t. I remember from back in the day the Ted Turner purchase of the MGM library. Many in the junk bond community were outraged that
wanted to piss away so much money buying a seemingly useless idle asset. Actually, though, he was right. The economic payoff he got in terms of cash ultimately generated was way above what he paid. One of the best decisions I made was to be among the first to sign up for his IPO. He didn’t give dividend discount presentation, but for those who really paid attention to what he was saying, it was clear he was following the framework perfectly.

It’s not the formula. It;'s the logical framework.