🧵Is the stock market being manipulated this year by some "agents? This is a question some #traders ask after looking at this smooth $SPX uptrend since early January this year. 1/n It is hard, or even impossible, to prove the market is manipulated because the manipulating agents are also market participants. But we can look at some statistics from the above chart. The mean daily return from 1945 has been 0.03% versus 0.18% in the last 60 trading days. 2/n

🧵You do not need fancy analysis for being bullish the stock market. There is so much noise and misunderstanding. All you need is a moving average and monthly prices. It has been so easy. Caveat emptor in numerous blog article. 1/n The "simple" 12-month moving average has delivered 9.2% annualized return (backtest) in $SPY ETF since inception with a beta of about 0.5, Max. DD is less that half of that of buy and hold at a cost of 100 basis points. Sharpe is 0.73. 2/npriceactionlab.com/Blog/2024/03/m…

🧵What is a hedge for a stock portfolio?

First, what is not a hedge.

1/n Bonds are not a hedge. 2022 was the proof. 2/n Gold is not a hedge. Gold is a catastrophe hedge subject to several conditions, but only in physical form.

🧵Sharpe ratio. Based on recent but also past discussion, it appears to me some traders do not understand the value of the Sharpe ratio. This is probably the most important ratio for evaluating trading strategy performance. Let me explain why. 1/n Given that alpha is hard to generate in practice, most strategies aim for better risk-adjusted returns. This is where the value of Sharpe comes in. I will first try to explain what a low Sharpe value means. 2/n

Happy July 4th! 🧵Correlation and cointegration.

The thread is a response to a recent discussion involving more dual-axis macro charts and their use in market forecasting. 1/n One of the charts showed the S&P 500 and some proprietary index and the claim was that a forecast can be made since the correlation appeared (visually) high. Obviously, the claims are bonkers. They reveal major misconceptions and statistics illiteracy. 2/n

🧵I hope this will be helpful to some people.

A maxim of trading strategy developers is that Type-II errors (missed discoveries) are preferable to Type-I errors (false discoveries). In other words, it is better to miss a good strategy than employ a bad one and lose money. This trade-off arises from the tests used to test strategies for robustness and significance and the fact that the future may be different from the past. Guarding against p-hacking increases Type-II errors and missed discoveries unless the power of the test is high.

Short🧵@rjparkerjr09 posts these interesting discussions but some raise red flags. The "forecasts are pointless" claim is a distracting issue. This claim is usually made by salespeople to appease investors who do not like market timing, or by some confused. 1/n

https://twitter.com/rjparkerjr09/status/1608189117690703872

The claim that trend-followers do not forecast anything has been debunked by many, but this article in the QUSMA blog offers elegant proof. qusma.com/2014/06/30/tre… 2/n

🧵CTAs vs. ETFs that track CTAs

There are some important operational differences. Certainly, ETFs that track CTAs, $DBMF for example, represent a disruption and a genius idea in the managed futures space due to lower fees but this does not come free of risks. 1/n There is no free lunch in financial markets. CTAs have flexibility in dealing with inflows, they can manage accounts separately under a master account, and this allows for controlling risk. This is important but not easy with ETFs 2/n

🧵Financial Twitter is a mix of many different types with diverging objectives and different backgrounds. The result is a high entropy environment with a low signal-to-noise ratio. In the best case, it is entertainment, and in the worst, misinformation. 1/n 1. Retail w/ experience.
2. Retail w/o experience.
3. Pro traders in various markets.
4. Small fund managers/CTAs.
5. Large fund managers/CTAs.
6. Sell-side analysts.
7. Derivatives pricing quants.
8. Financial journalists.
9. Recreational traders.
10. Macro analysts 2/n

🧵Most #traders do not realize or are unaware due to lack of education in the area, that when they enter a trade, they are basically betting on the presence of a price action anomaly, or market inefficiency/outlier. 1/n When there is no inefficiency present, the trades over the long term have zero expectation minus transaction cost, market inefficiencies provide a positive mean and skew to the P/L distribution. But it's not easy. 2/n

🧵Introductory thread about the definitions of probability and how they relate to trading and investing.

There are advanced definitions of probability and even versions with negative values but for the purpose of this thread, I'll talk about the basic level. 1. Axiomatic definition.
I. P(α) ≥ 0
II. The probability of a certain event equals 1
III. If two events α and β are mutually exclusive, then P(α+ β) = P(α) + P(β)

This serves as a foundation for developing a theory and proving theorems (along with some theorems from algebra).

Short thread 🧵Time to talk about some real quant work. 😂
Daily machine learning was challenging in the first part of the year. Features had a low signal-to-noise ratio. Adaptation started after may and machine learning has recovered. In the case of $SPY YTD return is 18%. 1/5 The features are calculated by DLPAL LS software. priceactionlab.com/Blog/dlpal-ls/
For single securities, we use the directional bias P-delta for timing. If Pdelta is > 0 the algo goes long, if P-delta < 0, it goes short. 2/5

🧵Twelve years ago I worked long hours every day for many months testing ML algos for trading and also later for the Numerai contest. Endless testing, scoring, and backtesting of maybe two dozen ML algos. 1/n I almost became a burnout. I realized that ML algos are essentially optimization schemes. No good generalizations. It's the features that count. If you think some algo from a library will make money in markets, you'll be disappointed after wasting time and maybe money. GIGO. 2/n

🧵There are some misconceptions by some trading book authors but also academics regarding non-stationarity in financial markets. Financial price series are non-stationary random processes and not even in the wide sense. All moments are random variables. 1/n Some people blame losses on non-stationarity because trading strategies have problems dealing with even a wide sense of stationarity. This is true. What these confused people don't realize, is that profits are also due to non-stationarity. Make a note of this. 2/n

Thread 🧵about the market.

1/n Family, friends and some followers ask me where I think the market is heading. Some are bulls, others are bears. It's a good question. Here is my answer. I have no freaking idea. $SPX $SPY $NDX $QQQ #ES_F #NQ_F 2/n Maybe I'm not as smart as some gurus who predict a rebound to new highs, or some others who think this is going way lower. Or, I have learned my lessons. This is what I mean.

Thread 🧵 CTAs are having a good year so far with Top 50 +12.8% YTD, according to BarclayHedge. But let's not forget that from 2004 to 2021, top 20 CTAs returned about 2.5% on the average.

"But, but, but, convexity, alternatives, etc." #commodities #CTA #trading $SPX This is another form of tail risk hedging but with higher intrinsic risk than using other tail risk protection method. The 80s and 90s aren't coming back, when CTAs using TF discipline profited from naïve TA traders in commodity futures. Nowadays there are no such fools en masse.

Thread about trading strategy and fund performance analysis. #tradingstrategy #funds

Advanced performance analysis is fine but nothing beats common sense. Below are some common sense ways of analyzing strategy and fund performance. 1. Annualized return

The strategy must offer an acceptable annualized return subject to constraints. Very low annualized return is undesirable but also high annualized return that comes at high risk is also undesirable and even more so.

In strategy development, false positives (Type-I) can cause losses but false negatives (Type-II) can cause missed profits. This problem cannot be solved quantitatively at a high level. The answer is experience and no algo has that (except if it's a sophisticated expert system.) A false positive is when you select a strategy that was the result of data snooping or over-optimization but the Sharpe, t-stat or p-value looked good. These may fail immediately but in some cases can remain profitable for many years due to luck of persisting market regimes.

Short thread. I see some charlatans and especially one of most notorious, which I won't name, arguing for high probability of rebound tomorrow in stocks. I posted a backtest yesterday with results from shorting at close of Friday if larger than 2% drop.

https://twitter.com/mikeharrisNY/status/1464696599834251268?s=20

These backtests have small samples and particular one shows, as expected, there is higher frequency of rebounds on Mondays after a large drop on Fridays, of around 60%. Now, frequency in historical data hardly translates to probability and also even then, 60% probability is low.

Thread. We consider the question. Why regulators in USA are not acting to eradicate the cryptocurrency and NFT pandemic? In other countries they have already taken action. The answer is as always complicated because everything in Western World has become so complicated there may be no answers to even basic questions: Complexity has increased beyond bounds.

Thread now that the market is closed. 🚀

Let's talk about going short after another permabear fund announced shut down.

But before that, I'll talk briefly about going to the casino. It's related in a way. I don't like casinos because I don't like to be around so many losers. The energy is bad. Casino players essentially go short the casino with tiny odds to win.

Last time I visited one was many years ago. I made windfall profits with a martingale system playing the roulette.