Sociopath Community

Data
Year Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
1997 3.45 2.15 1.89 2.03 2.25 2.20 2.19 2.49 2.88 3.07 3.01 2.35
1998 2.09 2.23 2.24 2.43 2.14 2.17 2.17 1.85 2.02 1.91 2.12 1.72
1999 1.85 1.77 1.79 2.15 2.26 2.30 2.31 2.80 2.55 2.73 2.37 2.36
2000 2.42 2.66 2.79 3.04 3.59 4.29 3.99 4.43 5.06 5.02 5.52 8.90
2001 8.17 5.61 5.23 5.19 4.19 3.72 3.11 2.97 2.19 2.46 2.34 2.30
2002 2.32 2.32 3.03 3.43 3.50 3.26 2.99 3.09 3.55 4.13 4.04 4.74
2003 5.43 7.71 5.93 5.26 5.81 5.82 5.03 4.99 4.62 4.63 4.47 6.13
2004 6.14 5.37 5.39 5.71 6.33 6.27 5.93 5.41 5.15 6.35 6.17 6.58
2005 6.15 6.14 6.96 7.16 6.47 7.18 7.63 9.53 11.75 13.42 10.30 13.05
2006 8.69 7.54 6.89 7.16 6.25 6.21 6.17 7.14 4.90 5.85 7.41 6.73
2007 6.55 8.00 7.11 7.60 7.64 7.35 6.22 6.22 6.08 6.74 7.10 7.11
2008 7.99 8.54 9.41 10.18 11.27 12.69 11.09 8.26 7.67 6.74 6.68 5.82
2009 5.24 4.52 3.96 3.50 3.83 3.80 3.38 3.14 2.99 4.01 3.66 5.35
2010 5.83 5.32 4.29 4.03 4.14 4.80 4.63 4.32 3.89 3.43 3.71 4.25
2011 4.49 4.09 3.97 4.24 4.31 4.54 4.42 4.06 3.90 3.57 3.24 3.17
2012 2.67 2.51 2.17 1.95 2.43 2.46 2.95 2.84 2.85 3.32 3.54 3.34
2013 3.33 3.33 3.81 4.17 4.04 3.83 3.62 3.43 3.62 3.68 3.64 4.24
2014 4.71 6.00 4.90 4.66 4.58 4.59 4.05 3.91 3.92 3.78 4.12 3.48
2015 2.99 2.87 2.83 2.61 2.85 2.78 2.84 2.77 2.66 2.34 2.09 1.93
2016 2.28 1.99 1.73 1.92 1.92 2.59 2.82 2.82 2.99 2.98 2.55 3.59
2017 3.30 2.85 2.88 3.10 3.15 2.98 2.98 2.90 2.98 2.88 3.01 2.82
2018 3.87 2.67 2.69 2.80 2.80 2.97 2.83 2.96 3.00 3.28 4.09 4.04
2019 3.11 2.69 2.95 2.65 2.64 2.40 2.37 2.22 2.56 2.33 2.65 2.22
2020 2.02 1.91 1.79 1.74 1.75 1.63 1.77 2.30 1.92 2.39 2.61 2.59
2021 2.71 5.35 2.62 2.66 2.91 3.26 3.84 4.07 5.16 5.51 5.05 3.76
2022 4.38 4.69 4.90 6.60 8.14 7.70 7.28 8.81 7.88 5.66 5.45 5.53
2023 3.27

Whats crazy about this data is that the price really has not changed since 1997. Sus man, real sus -_-. Has the price of energy really not changed in decades?

It's gone up and down and back to the same value as Jan 1997.

I'm starting to think this market is being manipulated.

look at the confidence interval, lol

LiYang said: 

 

look at the confidence interval, lol

 shitbulls always have the most egregious predictions xd

Today I made another 100 dollars trading natural gas!

https://energymarkettrader.blogspot.com/2023/02/22123-energy-market-trading-journal.html

What happened to muu fucking images.

Golden_Eagle said: 

AliceInWonderland said: 

LiYang said: 

I've been thinking about machine learning but haven't had the time to put into the research.

 

Forecasting on metal resource spot settlement price: New evidence from the machine learning model

"We found that LSTM-GRU and other models also perform well with strong robustness. Therefore, we believed that the LSTM hybrid model, especially the LSTM-GRU model, is suitable for analyzing the prediction of spot settlement price of metal minerals."

https://www.sciencedirect.com/science/article/abs/pii/S0301420723000685

 

Maybe a simple natural gas economic model with supply and demand inputs. Start simple and see what others have done.

Forecasting doesn't matter as it doesn't inform your actual PNL. 

You can be right about the behavior of a random variable X and still go bust if your pay off function f(x) doesn't align with what your forecasting. 

Moreover, given X in markets is a highly stochastic variable that is fat-tailed its very difficult to accurately forecast given the complexity of that variable. As such, almost all your focus should be towards your payoff function f(x) given its far simpler and doesn't rely on making predictions but instead odds and optimal betting. 

All the stupidity in statistics can really be summarized as the focus on values of functions instead of the functions of values. 

 There are indications that 2023 is going to be a time of recession. I might get them if I must.

Personally I find patterns very suggestive as to what's about to happen. 

I'm not sure what's a good hedge against recessions either, cause people will liquidate their assets after maxing out their credit cards.

 Right, maybe less demand for energy. If demand goes down prices should drop in the short term. Forecast for price drop, down. I think the fed. web site said the same trend.

At first look, It just seems this is another way of saying your confidence interval is large.

Or is fat tailed the description of the distribution of the returns? Not a normal distribution. They may even be multi-peaked and weighted off the norm.

"Forecasting errors are necessarily thin-tailed while market events and therefore payoffs are thick-tailed."

AliceInWonderland said: 

Liyang said:

You're making a lot of claims here with no references. Sorry, fake news wikipedia won't cut it. Where is the evidence of your claims?

Statistical Consequences of Fat Tails: Real World Preasymptotics, Epistemology, and Applications

I use this book as a reference a lot.

Read 2.2.22-2.2.23, 2.2.25, 3.4, 3.10-3.12, 11.1-11.3, and 12

You will find references for these sections in the back of the book, check their if you want to see empirical studies and proofs of derivations. 

I can't find a PDF online but Modelling Extremal Events by Embrechts, Kluppelberg, and Mikosch is an awesome book about the limitations of normal forecasting and more specifically how to make more robust ones a long with better alternatives. This book should be the bread and butter of finance but for some reason its not. More of a derivations book. 

If you want something interesting to read about that you can just google, look into the shortcomings of VAR forecasts and how they played into the 07-08 financial crises.

"Forecasting errors are necessarily thin-tailed while market events and therefore payoffs are thick-tailed."

Spatial Mind said: 

AliceInWonderland said: 

Spatial Mind said:

Your chart looks like Ryu's super move.

I'll see if I can find the data I was refering to, it'd be easier if I can find the video explination, it's very interesting. All pattern recognition based on economic performance, and we're reached the top.

Please do. 

A quick google returns only results related to street fighter.  

 

Thomas Kralow presented this yesterday, and it's what I was talking about. It basically shows data pertaining to how it's impossible to go from 3.4% Unemployment to 4.6% without first going into recession. This data goes back to the 40's I believe, which would mean this data has been consistent 11 consecutive times. 

The site has a paywall so skip it if you wish. The video has it the details

https://ecoinometrics.substack.com/p/ecoinometrics-the-federal-reserve

This video should be time stamped from 4:08

Watched the vid. 

I'm not convinced by this unemployment idea simply because of the scaling of those graphs. Looking at the total graph there are periods where it lasts one year, two years, and three years. And, in many instances the spike briefly reaches 6% only to fall quickly with no negative growth. When he shows that future projections of unemployment he says you can't go from 3.4% to 4.6% unemployment and not experience recession, yet there are examples in the historical data where similar things have happened without negative growth. 

A recession doesn't equal and contraction. Regardless, maybe we will see an economic recession but I am not convinced that this kind of treatment of the data is a clear indication. 

He hypothesizes that BTC and S&P are decoupling, and his reason is BTC move since December. The issue with this is that S&P has seen a similar move. The difference is that BTC on average has a much higher volatility than the S&P, so while S&P may only move ~20% over the same period as BTC 50% gain doesn't mean they are not 'coupled'. It just means that one is more volatile than the other. 

To measure this possible decoupling you could use something like Mutual Information to actually confirm the validity of said hypothesis. Just looking at a chart and its scale of movement isn't at all informative. 

If you use correlation to measure the two, the more traditional measurement of dependence (despite the fact that it doesn't measure dependence and its stupid to measure two nonlinear systems), they still have positive correlation. 

LiYang said: 

At first look, It just seems this is another way of saying your confidence interval is large.

This does have a lot to do with it. 

For instance you shared the following forecast, 

The value of this forecast isn't the forecast itself because you don't make money off of forecasts. You make money from trades, hence any real value from the forecast comes from the information that informs a trade. 

The valued information for large firms, banks, institutions, etc. has little to do with the actual projected plausible prices, it has everything to do with how this informs how much risk is reasonable to take on your books by deciding to make a trade. 

So if I am 95% confident in my price bounds and those bounds allow me to measure the risk of a decision, I better be really sure that 95% confidence is the actual confidence or I will be underestimating my risk and not betting optimally. Underestimating risk leads to you take on more risk than you should, which leads to betting larger than the actual optimal amount. In the long run that leads to ruin. 

In the graph I shared, 

The red dots are the optimal bets. Notice how dangerous betting beyond those points becomes. 

The major issue is that 99% of the time forecasts are built on cookie cutter statistical methods that underestimate risk, but work a large proportion of the time. Despite working often, they don't work exactly when you need them. And, the whole time you've been following them you've been over-levered. 

Or is fat tailed the description of the distribution of the returns? Not a normal distribution. They may even be multi-peaked and weighted off the norm.

Returns are typically the major focus and this is not different for standard forecasting techniques. 

The reason for this is that returns are stationary while prices, especially prices of securities, are non-stationary. That is returns have a distribution while prices do not, as such you can much more reliably say informative things about returns. 

Returns are indeed fat-tailed. Forecasts of returns are often thin-tailed unless you are incredibly careful. 

today i made 200 dollars trading natural gas!!!