A Conversation with Renaissance Technologies CEO Peter Brown | Hacker News Reader

A Conversation with Renaissance Technologies CEO Peter Brown

goldmansachs.com

I remember reading a reddit thread that detailed the broad outlines of one of Renaissance's trading strategies. From memory it was basically arbitraging indexes of certain groups stocks by buying/selling the underlying stocks using leverage when the index deviates from the price of the assets. They used predictive modelling from lots of data sources to determine when this would occur.

Does anyone have the link?

This us just simple index/etf arb. It is not about what they do there is certainly something in how they do it and it is most likely something common that they replicate in all their strategies.

There's nothing new in what you described. Saying "they arb and use lots of data" contains exactly as much information as saying "they buy low and sell high". The implementation details are everything.

The Reddit link I was hoping someone had contained specific symbols and detailed information about the strategy. I kept my summary general because I wanted to keep the summary as accurate as possible.

I hope someone answers with the link, I'm very curious myself.

The Reddit discussion on algo trading: https://reddit.com/r/algotrading/s/zYzV6XxR0p

Quora discussion from an informed source: https://www.quora.com/What-are-the-investment-strategies-of-...

Also talks about the barrier option they buy from investment banks for hedging purposes and the tax situations.

I recently read Reanaissance co-founder Jim Simon's biography [1]. It seems like they discovered some secret algorithmic sauce that nobody else has found. But that feels implausible. Would love to understand what really gave them their edge.

[1] https://www.amazon.com/Man-Who-Solved-Market-Revolution/dp/0...

It probably is the case that they did discover something (whether an algorithm or good business model via legal tax evasion). Imagine if einstein went into financial trading instead of physics. I think what regular people fail to understand, there are people out there with the intellect to make hundreds of millions. Meaning, they can understand society on an abstract level that makes good business opportunities obvious. Most will never interact with a single person like this in their whole life.

I think believing in this kind of determinism is harmful, mostly to the individual who holds the belief, but also at scale, when many fools hold this belief. It makes you ripe for exploitation and deceives you into being or staying a loser, "oh if only I had that arcane magic ability that he has...". Fortune, grit, and deception [1] make kings, not some kind of magic superhuman sixth sense.

[1] Deception also prevents the making of kings, as in the case of the man who believes magic is what he'll need to succeed, and it's this magic he doesn't have.

No its just the reality in a capitalist economy. Human beings are normally distributed, by definition there are people on the far right tail of ability. Its a numbers game.

Eh. I studied math at an elite university with a lot of people who ended up at these secretive quant hedge funds. They're undeniably brilliant, and as per GP also extremely motivated and hard working (brilliance alone doesn't cut it!).

They're also heavily dependent on collaboration with others, availability of data, the existence of the right tools, and so on. Unlike in comic books, raw intelligence alone doesn't get you very far. But put a bunch of very smart people together for a decade+ with a lot of money on their hands and you can definitely make magic happen.

Academic achievement doesnt mean someone can come up with new things.

Leading people in their fields tend to come up with or find new things.

Unlike in comic books, raw intelligence alone doesn't get you very far.

what about solo tech start-ups?

> Fortune, grit, and deception

Well yeah so now instead of a super intellect you need grit. Good luck with that if you're born a low energy lazy person. :-)

I don't think Einstein going into financial trading would have necessarily translated to him being incredibly successful as trader or building a trading firm. History is filled with stories of incredibly smart people ending up in finacial troubles.

e.g. - Newton's investment in South Sea Bubble - Long Term Capital Group with multiple noble laureates imploding so bad that it almost took down Wall Street in late 90s (When Genuis Failed is a fascinating read

Which reminds me of a quote attributed to Newton: "I can calculate the movement of the stars, but not the madness of men."

Great book! Seems like they had several secret-sauce break throughs.

Seems like if you find an edge, trading on it removes it so others cant see it and copy.

Also these strategies don't scale infinitely. Medallion Fund is closed to investors presumably because they can't deploy that much capital towards whatever the strategy with an edge is.

One interesting bit from the book was that the strategies did not revolve around speed. It's not high frequency trading, or at least wasn't, but much slower moves.

Strategies also decay. Every quant fund is constantly generating new algorithms, because the value of old algorithms drops over time as others discover them.

Strategies decay if others discover them. RenTech claims to have a strategy that completely removes a certain arbitrage from prices if executed correctly. It would still show up in volume, but you can't extract the critical parameters from anonymized volume.

This would check out as a winning strategy. If you figure how to remove this arbitrage out before anyone else, then you nobody else can replicate your strategies. Since they were early in the quant game, this seems plausible (though other things mentioned here also seem plausible -- insider trading, lots of brilliant people, etc).

surprising how after 30 years no one else has figured out how to to this

Great book indeed. My favourite quote from it is: "The name of the game is not to always be right, but to be right often enough".

Bob Mercer said that they were right 50.75% of the time.

Interesting to note that that's larger than the house edge in video poker (which is admittedly one of the lower edge games).

You can build a massive city in the middle of a desert with edges well under 5%.

They probably found many small things and combined them extremely well. In quant trading, if you've got a lot of excellent researchers, chances are that you make better choices in lots of "micro" decisions that add up to a superior outcome.

Old college buddy switches around between fintech executive, State Department fixer, and finance journo. I asked him the same question about RT back in 2018, and his[1] reply was memorable:

"Secret sauce is always 'insider trading'"

He expanded on that a bit to also include buying up all the critical fiber nodes between the trading datacenters, and also laying down a new kind of fiber side by side with the old ones that's a teensy bit faster, and most critically, it knows how much faster it is. But the really big breaks - again, according to this buddy, who of course none of you know, so this has zero credibility - always come back to what's basically spies.

I don't think he's correct because of simple economics - spies and insider trades aren't going to be a smooth curve, and your investors will be pissed if one year you give them 5000% and the next 12%. Sure, you can move the money around to make a curve, but will your board let you do that? But yknow what? I don't work in the industry, so benefit of the doubt goes to the buddy who does. I'm not sure I could deal with working in the industry, given that the financial industry's done more to break this country than any other singular force. I would overdose in irony.

[1] He's a very cynical man

They mostly hire Physics Ph.D.s in the same close range of fields. As much as I find the insider trading narrative to be more plausible than any secret sauce, I think their hiring speaks against that.

Eh, I've worked with Math and Physics Ph.D.s throughout my career (and I am one myself). Those qualifications are definitely insufficient evidence on their own that you could beat the market. Not to say that hiring isn't their secret sauce, but you'd have to hire the _right_ Physics Ph.D.s, whatever that means.

I'm not saying that they make money because of Physics Ph.D.s. I'm saying it would be kind of useless to hire those specific people only to end up doing insider trades the whole time.

> As much as I find the insider trading narrative to be more plausible than any secret sauce

I would be interested in hearing why you find it more plausible.

I'll start. I find it less plausible because you actually go to jail for that.

If it's a secret sauce that is based on public information (aka not insider trading), then it's unlikely that no one else but this one company has discovered it. It's just extremely unlikely. Plus, plenty of "godlike investors" were clearly willing to do things that landed them in jail (SBF, Madoff, etc).

Hiring as an edge makes a lot more sense. Quant trading is often not about a single strategy but about the sum of many small ones. And other quants that have been taking similar hiring approaches also had incredible success (eg. Jane Street, Citadel, Jump Trading, Optiver, etc). RenTec just has a reputation of being "the best", which in turn makes the best people more likely to join them, which again in turn makes RenTec better, like a vicious cycle.

I also want to mention that in some years, there's a massive discrepancy between RenTec's internal Medallion fund (+76% in 2020), and its external funds (-19% to -31% in 2020). Not saying there is something to it (they never claimed using the same strategies), but it does seem at least a bit suspicious.

> Hiring as an edge makes a lot more sense. Quant trading is often not about a single strategy but about the sum of many small ones. And other quants that have been taking similar hiring approaches also had incredible success (eg. Jane Street, Citadel, Jump Trading, Optiver, etc). RenTec just has a reputation of being "the best", which in turn makes the best people more likely to join them, which again in turn makes RenTec better, like a vicious cycle.

Weren't you meant to make the case for insider trading? This isn't insider trading.

> I also want to mention that in some years, there's a massive discrepancy between RenTec's internal Medallion fund (+76% in 2020), and its external funds (-19% to -31% in 2020). Not saying there is something to it (they never claimed using the same strategies), but it does seem at least a bit suspicious.

Are you telling me that the insiders keep the good stuff for themselves and the outsiders get fucked? I don't find that suspicious at all.

I'm supporting dna_polymerase's point, that RenTec's edge is much more likely to be hiring smart people than insider trading and more likely to be insider trading than some secret special trading algorithm that no one else managed to come up with in 20 years.

To my knowledge, RenTec aren't market makers, they supposedly run directional bets. It is highly unlikely to do that with their size of funds ($130B) for that long (>20 years) with that success rate (>60% p.a.)

Without signing on to the belief that RennTech is generating returns from insider trading, I will observe that hiring a bunch of Physics/Math PhDs would be a reasonable cover story and therefore wise thing to do, even if your core returns came from insider trading.

In addition, it could be that the PhDs' technical research and work is about how to structure a portfolio strategy such that it can profit from insider trading when the opportunity arises without triggering any SEC alerts.

For example, maintain a portfolio across as wide an asset base as possible, while trading each asset frequently enough that any insider trade inserted into the mix looks random, or at least plausibly deniable, and thus a low-odds case for the SEC to prosecute.

This doesn’t seem to me at all cynical, since members of congress engage in insider trading without consequence.

They were successful when they were trading at lower frequencies and they hired people who implemented successful strategies that are pretty well known at this point to come work for them. People like Ed Thorpe had similar Sharpe ratios with less leverage so I think their returns are entirely realistic. It also took them about 10 years to get consistently profitable (under different names).

The low-latency networks you're alluding to exist but there is nothing nefarious about them. They aren't insider trading they are just getting market information faster than almost all other participants. I also don't think RenTech does this, it's more of an HFT strategy, and the HFT industry as a whole makes a few billion in profits per year, a lot of money sure, but it's not a huge amount when you consider how many transactions they are involved in.

Also, while the financial industry has caused some major blowups and bad actors often go unpunished it's a crucial part of any developed country and enables almost all big projects to get funded.

insider trading cannot be scaled up as well and smoothly as what Ren Tec does

I agree that insider trading is an unlikely explanation for the reason you mention.

It could be that an early algorithmic advantage compounded into an incumbency advantage: better research tools; better data; better pay for better scientists.

The 'secret sauce' may have started out as novel discoveries and, over time, become 'just being RenTec'.

Buying up fiber wouldn’t be a way to make money in this space. The space is dominated by microwave transmission links as they’re considerably lower latency than fiber.

Yeah, he was talking on old info, probably because when I was hanging out with him he'd just gotten back from five years in <LOCATION_REDACTED>. Microwave, shortwave, and that weird dual-mode laser AOptix thing have been the rule for a bit, although I don't know if the laser thing ever really worked. Maybe enhancing the other things. Probably by now they have their own satellites.

Microwave has been the standard for about 13 years now. Laser hasn’t panned out as I understand. Satellite wouldn’t help much because the bounce up is longer than just shooting in a near straight line with repeaters for long haul.

I had a high school friend with a relative who was a reasonably-successful hedge fund manager ($500M AUM in the early 2000s) - nothing crazy, no insane quants, just a pretty traditional (for the time) fund, and in moments of candor (usually assisted by THC), the relative basically said the exact same thing:

Most of the alpha in the traditional hedge fund world often just boiled down to getting away with insider trading. It was so rampant that it was just part of the game. The stuff we all heard about in the news was just the particularly egregious stuff the SEC HAD to blow up. But if the SEC prosecuted every instance of anything even remotely insider-y, there would have been no hedge funds left.

(Now to be fair: I'm not really speaking to Renaissance - just to the general industy in the early 2000s. I think other commenters have made compelling points that Renaissance may very well have some actual special sauce.)

I was friends with some hedge fund guys in the 2000s. The way they presented it was that some funds have good technology but in the end there is a ton of insider information going around. It may not be strictly illegal (I don’t know) but people talk to each other and sometimes work with each other. They also have access to company executives the regular investors don’t have.

I don’t think it’s possible to get sustained outstanding results without being part of the insider circle. As far as I know one of Warren Buffett’s close relatives was in Congress which probably also gave him access to people most regular citizens won’t have.

His dad was a congressman but he died in 1964

I also read the book and liked it a lot. Based on the extensive focus the book gave on their data collection (e.g. going into local state offices to digitize records), pipeline and academic rigor, i would think it's just looking for correlations and focus on medium term trades (minutes)

It's easy enough to do, but it requires a strong culture of honesty to not trick yourself. For instance you can regress performance based on 32 different lags on another variable and you'll almost certainly find spurious correlations. But in a more comprehensive framework you can adjust your score to account for such a search.

The people in the book state that they see the data no more than numbers and don't even try to come up with a story. And they can only do so because they trust their system to handle spurious correlations

Early on, they did find secret algorithmic sauce. By now, they are so large that their edge largely comes from their ability to see stuff before anyone else.

I think there's actually enough in the book to get a reasonable idea of what's going on, as well as an explanation for why you can't just do this yourself.

1) Trading PnL is very steadily profitable for their main fund. It must be trading a lot and winning due to the law of large numbers.

2) Their futures fund has normal returns. It looks like a lot of other CTAs, which also are not trading on subsecond time scales.

Chances are they have some mix of liquidity provision and arb strategies in Medallion, which normally is not a hedge fund, but they started in an era where that was what people did, so they ended up with a hedge fund and gaining reputation as such. This is also why we know anything at all about the returns. Nowadays you would have a prop trading firm and nobody would know, but people who know, they know the returns are quite steady.

In fact IIRC the book mentions specific people they'd come across in the early days trading certain kinds of strats that are known to be profitable enough to beat your average hedge fund. My guess is they found a few of these and refined them.

For the CTA strat, they run a hedge fund because that's the appropriate vehicle for a thing with a sharpe/information ratio around the 1-2 range.

The thing that really makes it work for them is probably incumbency. They've been around forever, people in the business want do business with them, and they know where all the landmines are, both in regards to trading strategies and legal/business framework. Even if you knew the strategies you would struggle to set up all the infra, contracts, relationships, and so on.

Fascinatingly, Jim Simon's philanthropic org, Simons Foundation, is also what created and funds the well known Quanta Magazine.

Also the (formerly known as) Mathematical Sciences Research Institute at Berkeley is named after him after a large grant. It's featured frequently on the Numberphile YouTube channel (as they fund Brady in some capacity).

The value per dollar in sponsoring educational content on YouTube is incredible. Imagine how many people are learning from channels like Numberphile, 3Blue1Brown, Steve Mould, etc.

I found a new-to-me grad student guy, with decent video skills, got 240k views on a certain paper review video. wild days..

I'm not sure it's generally understood what a serious academic research mathematician he is/was. He taught at top programs, won awards - https://en.wikipedia.org/wiki/Jim_Simons_(mathematician)

I interviewed with them way back in 2006, but regrettably did not get the job. I was also interviewing with a lot of other banks and hedge funds at the time but these guys really stood out. I was impressed just how friendly and seemingly happy everyone was. It felt more like a university research lab than a trading firm. Everyone had their own private office and they ate lunch together at the cafeteria. One guy told me with eyes wide open: "If you get this job you'll never want to leave".

Interview questions were standard fare, the usual brain teasers; one guy was really obsessed with calculating covariance matrices. They had a wood-panelled library with all the scientific journals where I spent the time between interviews. The gentleman who was guiding me through the interview process told me how their brokers were constantly front-running their trades so a big issue for them was keeping their market impact minimal.

During the interview I offered some potential ideas for alpha: having done some research on these guys before the interview I knew they had experts on voice recognition so I suggested they could put a microphone in the trading floor and gauge sentiment via spectral analysis. They really liked the idea. Trading floors no longer exist so this alpha is long gone, if it ever worked at all!

But alas - their alpha is simply skill and hard work - there are no shenanigans or insider trading going on there. It's just time-series regression done perfectly. I bet at the time they were not using anything fancier than linear regression. But considering how large their data team was they probably had access to datasets (even today) which give them tremendous edge. Data is oil in this field.

The book "MindF*ck" [0] briefly describes a project within RenTec to build a statistical behavioural model of every individual citizen the USA, fed by enormous public and private databases. From this behaviour they could anticipate consumer patterns which could predict the markets way ahead of anyone else. The insane level of ambition in these projects gives you a clue as to the level of technology they potentially might have.

[0] https://www.goodreads.com/en/book/show/52269471

(Audio)

There's a transcript pdf linked at the bottom of the page after the small print.

I read the transcript first and then checked out the audio to get a flavor of it. Maybe it's just because I read it first, but it almost sounds like Brown is reading his answers from a sheet. And/or there's some automatic removal of pauses.

I thought the term "transducers" (instead of transformers) might have been a transciption error, but that's really what he says.

It's not that hard to find good quant strategies. A strategy that goes long QQQ at the market open and shorts Bitcoin, equal allocation, and then closes both position at closing bell, has posted very good returns all year

very easy $

I imagine this could be scaled up with other strategies such as what Renaissance Technologies is doing. I am running this now and had my biggest year ever

> I imagine this could be scaled up

Likely a false assumption. The flagship Medallion Fund supposedly has $10B AUM. Many strategies that work in insignificant volume (such that your actions don't affect the broader market) won't work once your size becomes such that you entering and exiting the market moves it significantly.

> I am running this now and had my biggest year ever

Kudos! Are you always looking for other strategies, since I'd imagine the lifespan of any individual strategy isn't that long?

I don't mean the Bitcoin trading strategy scaled up. I mean the idea behind it scaled up to many asset classes.

Reply on news.ycombinator.com