Hacker News Reader

Comment by pyrrhotech | Hacker News Reader

IMO it's easier and less work to outperform the indices with a combination of passively holding ETFs like VTI / VOO and strategic hedging based on quantitative leading indicators of economic trouble rather than trying to pick individual stocks. It's also more tax-efficient since you can continue to defer gains in the held ETFs while hedging with Section 1256 contract futures or options instead of constantly trading in and out of individual stocks and being hit with a lot of capital gains tax.

My system is down -1.58% this year compared to -17.39% for the SPX for a nearly 16% outperformance margin. I've documented the results and information about it here: https://grizzlybulls.com/models/vix-ta-macro-mp-extreme

View on HN

I'm skeptical:

1) Your performance chart starts right after the 2009 bear market and includes a huge bull market run up until the current slowdown where it then starts to match SPX. The straight line up until the drop starting in 2021 suggests that your model may not perform so well in the future.

2) You don't include the penalty from all the short term capital gains taxes you're generating.

3) If you can really generate those returns you wouldn't need to sell market timing signals.

Healthy skepticism is warranted! I've answered several of these in previous HN comments such as https://news.ycombinator.com/threads?id=pyrrhotech&next=3012... and https://news.ycombinator.com/threads?id=pyrrhotech&next=3012... but in brief:

1. Vix Futures are a huge part of all the VIX based models. Vix Futures came out in 2004 but the earliest intraday per-contract data publicly available is in 2009 which is needed to calculate the futures curve.

2. Slippage is included in the modeled returns but taxes are not, as mentioned in the tooltip. Taxes are not included because they vary widely by location, account type and implementation method. I've also written at large about how to minimize taxes on the blog. As a conservative estimate, feel free to multiply returns by 0.75x to get equivalent after-tax CAGR, but in most cases you'd beat this in real life.

3. Correct, and I make no guarantee that they will always be available. As of now, selling access to them in no way negatively impacts my own returns. I've met a lot of people in this journey and I enjoy leading a community and making an impact. I also make a significant amount of MRR that has grown substantially from the start of the year that gets reinvested into the models in my own account. I consider Grizzly Bulls to be a win-win alternative to the hedge fund industry for those interested in alternative investments. We've only ever had one Platinum member cancel, and given the model's underperformance in Q1 it was understandable. Since Q1, the model has been crushing the market, and it makes me happy to see our customers happy as well. Finally, as I mention several places, no one should ever expect any model to achieve 100% of its backtest performance, but there's enough leeway in the performance and drawdown figures to underperform the backtest and still generate substantial alpha, which has been my experience running them live for over 2 years now.

> Vix Futures came out in 2004 but the earliest intraday per-contract data publicly available is in 2009

According to https://firstratedata.com/i/futures/VX, VIX intraday is available starting 5-Aug-2008. But your chart starts April 2009, which is the exact bottom of the bear market.

Right, but as it mentions there, "Continuous data is from 5-Aug-2008 , Individual contracts start with VXF09 (March 2009)". To calculate the Vix Futures curve, our models need the individual contract intraday breakdown not just the intraday data for the front-month contract. I'm sure this data exists somewhere but I haven't found anyone willing to sell it to me. If you do, please let me know!

I can assure you that there's no conspiracy to start the models right at the start of the last bull market, and I'd be elated to get my hands on the data for this as well as a couple other indicators that don't stretch back as far. Though there is no guarantee, market volatility, bear markets and general downturns are actually the conditions that enable the models to outperform. During buy signals which can be quite lengthy in periods of low volatility during bull markets, the returns are exactly equivalent to buy & hold.

You must just be really lucky then because I see this at the top of one of the posts you shared (https://news.ycombinator.com/threads?id=pyrrhotech&next=3012...):

I've traded them with my own capital successfully since April, 2020, and I've averaged 75%+ annual return

April 2020, the exact bottom of the covid bear.

Not saying you're lying, and I don't know enough to be smart about the subject and I haven't read through your material, but it seems like a magic trick...

Your performance chart is misleading, I think, because it looks like it's showing hypothetical growth of an investment made in 2009 like we would normally expect from any other fund performance chart. Your site event says "$100,000 invested in the model at inception would have grown to $39,196,370". But your model was not investable until very recently. So it's essentially untested. The real performance numbers aren't anywhere close to that.

For this reason I think it would be fair call this out and to distinguish between the two types of trades shown on the chart: backward vs forward looking. Of course everyone would expect the backward looking trades to be impressive. The forward looking trades have already shown signs of weakness. It will be interesting to see how it does in the next few years.

Bottom line: this seems no different to me than someone producing a model that uses Jim Cramer recommendations as input to predict stock moves, fits it to produce near perfect historical trades", and then promotes their 60% CAGR as if it was real and actionable.

Good luck to you anyway.

PSA: Outside of stat arb and heavily leveraged market making with teams of PhDs, 60% annualized CAGR doesn't exist. In the real world, anything above 10% over the long-term (in-sample over 50 years), is extremely likely to be spurious and won't repeat.

Is this essentially a tuned version of the whole "Shannon's Demon" a.k.a martingale system that makes a run at the HN front page every six months or so?

This is different, no martingale here and no leverage used at all in obtaining the returns, the models are always either hedged (effectively in cash or 0% long) or unhedged (100% long). This is also not a high frequency trading system. Most models have an average holding period of about 2-4 weeks, though with high variation depending on market volatility levels (has been much shorter than that this year). Each detail page gives a brief, high-level overview of some of the indicators the model uses to detect trend strength, trend reversals and mean reversion opportunities that ultimately produce a signal change.

People often think beating the market is impossible, and for an unaided human, it may be. But just like Magnus Carlsen cannot come close to defeating Stockfish at chess, sophisticated algorithms that analyze far more data than possible for a human are on an entirely different playing field. The ironic thing is the growth in the belief in the EMH and passive investing has actually made beating the market easier in recent years as there's simply much less competition and more money buying no matter what the economic conditions or monetary policy are to exploit. I've written a blog post on my take in depth on the subject here: https://grizzlybulls.com/blog/time-in-the-market-vs-timing-t...

I view the Shannon algo going long dollar or long on a stock as a form of martingale, which is open for debate, but I was actually asking because I got slightly obsessed with disproving that it worked IRL, and started just messing with it, which got way past that and ended with me building my own model... and it seems to have interesting similarities to yours (superficially, at least). I've only recently put into practice. Purely because of my own risk tolerance, it's also long term, not intra-day, also not leveraged, and also only going long. One choice I made (for aesthetic reasons?) was to not use big data or anything beyond basic price histories, and to not use a neural net. I wanted something with simple, reproducible signal rules a human can understand, even at the expense of profit. I ran it through my own symbolic regression platform just to validate that I was onto something. I thought about posting it here and/or making it a paid service but it seemed to me (as someone said) that if anything like this makes enough money I wouldn't need to sell it, and would be better keeping it secret... plus I used to run a sports betting analysis site that was too much trouble and made me feel bad when predictions were off. As it is now, I'll only have to apologize to a couple of friends on a private email chain. Officially the target is 20% over S&P per year, but unofficially the results I'm seeing in the sim are 100% to 1500%.

I saw something it doesn't appear is present in your models, and I extrapolated from it into a program that chases that thing. Nonetheless it definitely seems like we've worked in a similar direction, as one potential application of my model is a simple long stock vs dollar trigger. And you've probably gleaned other insights I haven't. Offhand, would you be open to getting in touch and swapping strategies / pitfalls?

Very interesting, I'm glad to hear you are also building your own models and having success! I'm pretty tight-lipped about the inner-workings of mine, as unlike selling access to the generated signals, sharing details about the internals actually would hurt my edge and personal returns. That's why I've left the details pages intentionally high level and vague, in fact I think I may have erred on the side of too much detail already. I'm very impressed that you are seeing triple digit returns in the backtest of your model though without using leverage and long-only. Using long/short or modest leverage puts our platinum model into triple digits, but also increases the max drawdown and standard deviation of returns. I leave it up to the clients whether to use a riskier implementation such as this.

How long have you been running it live, and how far back does your backtest data run? How closely are you seeing real returns match backtest? After running live for over 2 years, I've seen up to 6 month periods where the model outperformed the backtest, but overall the models have underperformed to varying degrees as expected. Though I have spent hundreds of hours dedicated to removing noise and overfit bias, I've come to the conclusion it isn't possible to 100% remove it, which combined with expected market regime changes where legitimate patterns become unprofitable means that in the long run every model is essentially guaranteed to underperform its backtest, the question becomes simply to what degree. However, with a strong enough model, even underperforming the backtest can lead to substantial alpha which is what I and our members have experienced so far this year.

I've only been running it live for a couple months, so it's too soon to tell! Generally it wouldn't be expected to show any signals for the first 6 months or so. Additionally it's expected to weaken over time and need to be reset; 3 years seems about the optimal horizon.

The testing data I'm using goes back to Y2K. Overfitting is definitely a concern. I've proven that a delta that worked one year doesn't work the next, or what worked for one five-year span doesn't work on the next five. But what's interesting is that the local delta mean can be broadly broken into distinct eras by looking at sliding time windows. e.g. From 1998-2008, the best overall fit scales down well to each year. 2008-2012, 2012-2020, 2020-now... the last one is the great unknown, of course, and too early to call if I've gotten close to it. Any test that covers the covid disruption or the 2008 crisis has a nexus that skews future results to the high side, so I feel those can't be trusted. If I can beat the market by 20% in the first year, I'll be happy.

I'll keep an eye on your progress. Best of luck!

Reply on news.ycombinator.com