The problem with predicting absolute levels, is that there is a game theoretic aspect which undermines any mathematical trading strategy as soon as it is public. optimal game theory trading strategies don’t produce great results, and they are relatively trivial to identify. Instead strong profits in market long/short macro positions are mostly created by information advantages, which don’t really make for interesting Kaggle competitions. For example, big profits in macro trading have historically been consistently achieved by front running customer orders, by building timing advantages on top of trading infrastructure, by funding research analysts that inspect operations on the ground, by lobbying for regulations that change market directions and so on.
It’s very hard to tell if a best performing hedge funds that doesn’t have an unfair advantage, that declares its only using quantitative strategies, is in fact just a statistical anomaly with a hollow narrative.
>> The problem with predicting absolute levels, is that there is a game theoretic aspect which undermines any mathematical trading strategy as soon as it is public.
I took finance in Business school, coming from doing a lot of statistical analysis in a research lab. I hated my finance professors and there pseudo science. Pricing formulas work great until they don't. The problem is when they don't, they really don't, in a catastrophic way. Read "When Genius Failed." Real traders know this. But some economists and finance professors act like these mathematical models are describing a predictable physical phenomena.
LTCM in being overly leveraged, relied on other market participants to maintain short term price alignment, which meant it was not arbitrage. Salomon's reduced its role as market-maker, maintaining short term price alignment, which increased short term price anomalies, and thus increased LTCM's vulnerability. The Asian financial crisis increased the frequency and extent of those pricing anomalies, and the subsequent Russian Default crisis did the same. Margin calls were made on LTCM that it couldn't cover, forcing them to close out of their positions at very unprofitable times of the trade strategy.
So I don't think "pseudo-science" is a great description for what those B-School profs are teaching. Rather the pricing formulae are just the beginnings of the financial theory you need to run arbitrage strategies, but they are not sufficient. You need to augment them with a broader picture of market dynamics and capital management, just like you'd need to learn about financial law, financial market technology, and a bunch of other stuff to run a successful market-making desk.
From what I've read elsewhere, if you had managed to hold an LTCM replica portfolio until expiry, you would have actually made money.
https://www.wiley.com/en-us/Paul+Wilmott+Introduces+Quantita...
Start by reading/running notebooks ("kernels") and winners solutions from this and previous competitions with tabular/time series data, building models and then grid searching, stacking, bagging, boosting 'em up. Here's a bit aged but still useful guide how pros do it: https://mlwave.com/kaggle-ensembling-guide/
If you want something online, check out https://www.quantopian.com/
The reason why relative valuation models are more effective is the same reason why most sports betting models use current odds as an input. Prices contain information but, in my experience, these methods aren't totally effective because they often miss important information about the company itself (big price moves happen because relative valuations are wrong). Value or quality appears to do fundamental work but is often woefully blind (for example, there are proven accounting issues with value strategies...does your average quant understand this? No. Have they ever read a set of accounts? No. They have no hope. None.)
Just imo, I think quant strategies are almost totally worthless beyond liquidity provision (even a strategy like front-running news in FX...humans do this better, and I know people who are still making tons of money doing this). I think there is massive value in that mode of analysis but the people who make the most are always going to be people who know the fundamentals better (I think firms like Marshall Wace that are doing this synthesis will move ahead) because that information is often not in the price at all.
I'm not sure about the rest of your comment, but this is mathematically the correct reason why we don't predict absolute price levels.
And the reason why this is mathematically the correct reason is because for a non-stationary process, when you predict into the future, the variance tends to infinity which means taking expectation on any statistical model is useless since the variance is ridiculously wide
This isn't true if you're fast enough. Everybody knows how to do arbitrage but it's still extremely profitable if you're faster than everyone else. HFTs are consistently more profitable than slow trading firms, earning 40%+ returns, it's just a much more capacity-constrained form of trading so absolute returns are lower.
But yeah, it might not be that useful for the funds themselves.
But isn't prediction an inherent part of valuation?
Doesn't need to be. You have wealth you have to store somehwere because it is not being spent now. Can store it as equities, debt (including a bank account), Paper cash money under the mattress, gold bars, BTC, etc etc.
Let's say you have the list of potential places to store it and can get a price/value ratio but with a common term of "X" in the value
- A = 3.1/X
- B = 2.8/X
- C = 1.9/X
- D = 4.2/X
Yes, this is a massive simplification as it ignores diversification benefits and so on but there's the basics of the usefulness of valuation where you don't and can't know the value.
"prediction of the market" here means "Do you think the S&P500 is going up or down this year?" Warren Buffet completely ignores this, takes no view on it. Timing the market is very hard. Assuming the market will reflect the economy and the economy will grow because people work hard, are smart, we invent new things and population increases, then picking the best of what's available is the normal value investors route, a la Buffet as the most famous practioner.
Burton Malkiel's "A Random Walk Down Wall St." is still, IMHO the best summary of all the theory and practise of investing - excluding the quant funds like Renaissance. Maybe he has an update for them? Maybe it's also worth reading if it exists or maybe not. Don't know.
No idea if renaissance time the market or what.
edit: punctuation and the Malkiel recommendation.
Kaggle encourages a domain agnostic approach to modeling, in the sense that participants use sophisticated machine learning and statistical methods but typically have no domain expertise in the underlying data. This kind of approach to finance has historically performed poorly. [1]
Good quantitative trading is usually backed by a strong fundamental thesis and an interpretable model, which is obtained by cross-pollinating sophisticated math and statistics with domain expertise in some part of finance. That domain expertise might be in different kinds of assets, liquidity or market microstructure, but it's there.
$100k is cheap for Jane Street. If nothing else they have a new recruiting pipeline of people with demonstrable machine learning skills.
______________
1. I would also say this is a poor way to approach statistical analysis in most domains, and usually leads to spurious or overfit results. But the idea that you can just run a model and find patterns in pricing data is especially attractive and insidious.
Yes this is accurate and put very well. This is so much the case that if you have a strong background understanding of the field, the ML part can actually be picked up quite quickly or contributed by someone else. There are a few notable users who are both domain and ML experts and they tend to absolutely clean up in their field. I'm thinking of a couple of med students in particular who are formidable in every medical imaging competition.
I recently read 'The man who solved the market', about Jim Simons and Renaissance Capital. The way the book tells it, looking for patterns without seeking domain expertise (e.g. ignoring fundamental valuation of equities) is exactly what Renaissance did, and it worked out very well.
But all of this has to be utilized in the context of the data. The reality is that you're not going to develop a sophisticated options trading strategy without a strong understanding of what an option (and more generally, a derivative) is. You can't develop a viable statistical arbitrage strategy just by treating market microstructure as a blackbox signal to be solved with e.g. Fourier analysis. You can certainly find an edge in using fundamentally superior methods of analysis, but you still need to know what that data represents in the context of the market.
Don't be fooled: people working at firms like RenTech have a strong understanding of the underlying finance. It's just that they learned it on the job, because the ethos at these firms is that learning fundamental theory in math and statistics is harder than learning fundamental theory in finance. You don't have to take my word for it though. Read about one of the few strategies of RenTech's which has been publicized: https://www.bloomberg.com/opinion/articles/2014-07-22/senate.... Deutsche and RenTech didn't team up on this strategy (to fantastic success) by treating basket options as some kind of blackbox abstraction devoid of delta, gamma, theta and vega.
(Also, even if I agree it was purely intended for tax avoidance, I don't understand why you think that would obviate having to understand how the options work intimately well).
https://www.rentec.com/Careers.action?computerProgrammer=tru...
They look for programmers with knowledge of Tax and Risk Management.
However, RenTech probably has the cleanest data warehouse of any firm. It was revealed that they have PHDs who do nothing but sort data into databases.
I've only taken a quick look at the data, but the problem doesn't seem to be focused on their core competencies, but instead is much more general.
How can you tell? All the features are completely anonymized.
Admittedly, this challenge far oversimplifies the depth of the quantitative problems Jane Streeters work on daily, and Jane Street is happy with the performance of its existing trading model for this particular question.
I'd caution the HN crowd not to expect production-level quant models out of this, like I'm seeing some doing in the comments already. Kagglers are excellent machine learning practitioners and the models that come out of many competitions are top-notch stuff, often making their way into research papers. But this is a short competition on limited data in a non-real-world scenario. The winning models will be very interesting educational exercises and probably wonderful recruiting material for Jane Street, but won't be the underpinnings of a new fund.
That said, I can't wait to see what comes out of this one. It ticks all of my competitive boxes :)