/biz/ - Business & Finance

[Anonymous No.55338073]

has anyone here rigorously studied statistics and is now using it in trading?

[Anonymous No.55338075]

claim $700 in tokens here - loyatly.net

[Anonymous No.55338079]

>>55338073
not rigorously, more like i think of things to look up on wikipedia and tradingview script search, but yeah. its invaluable

[Anonymous No.55338091]

>>55338079
not only is the distribution important, but when the signal goes above or below the ideal, that is also significant as it reveals hidden ledges that are used in some smart nerds calculations

[Anonymous No.55338161]

isnt this more important for the volume chart than the price action chart?

[Anonymous No.55338177]

>>55338073
I do. Obviously it's critical in the extremely rational and efficient crypto market. You don't really need it to trade stocks though.

[Anonymous No.55338207]

>>55338161

[Anonymous No.55338227]

>>55338073
>no kurtosis
imbecile.

[Anonymous No.55338252]

>>55338177
have you taken any particular course / textbook for self study?
i'm currently looking at harvard's STAT110X on edX
https://www.edx.org/course/introduction-to-probability-0

[Anonymous No.55338411]

>>55338073
Did a lot of stats in my PhD, Bayes, reinforcement learning, you name it, but I just DCA into Bitcoin regularly.

[Anonymous No.55338895]

>>55338073
Picked up a couple of stats units in my data science degree. Best decision I ever made and learned a lot. Stats is easy, probability is a bitch.

[Anonymous No.55338958]

>>55338411
>>55338895
which stats textbooks did you guys use?
any good recommendations for self study?

>>55338895
have you read any quant books? seems like a good application of stats knowledge

[Anonymous No.55339545]

>>55338073
Where's the hypergeometric distr? :(

[Anonymous No.55339559]

>>55338958
Self study will take at least a year, likely longer. Those grad courses are theoretical, so it'll take time until you can finally use it to trade.

Look up the most important stats/prob concepts and trade as much as you can.

[Anonymous No.55339572]

stats is useless. No predictive power at all. I always love when the ''i love science'' atheist crowd keeps saying statistics brings knowledge.


science is 100% statistics, ie rationalism, which is why there is no truth in science. Rationalism is not compatible with empiricism. It is rationalists who try hard to merge their rationalism with empiricism and the end result is what they call science but there is no empiricism in this.
It's like a redefinition of empiricism by scientists.
Scientists are not empiricists. An empiricist does not run an 'experiment' and even less a 'thought experiment. Running an experiment already is part of rationalism.

The whole secular infatuation over empiricism started by the self proclaimed enlightened secular rationalists who hated the christian rationalists larping as greek philosophers.

The secular intellectuals hated the endless christian speculations that they saw as unverifiable about god and the use of logic from aristotle. So they said they could speculate on nature and with their verification meme, their rationalist system about the world would be the truth.
So in order to be separate from the christian clerics, the secular rationalists said they were empiricist. Their big idea was the empirical proof which is completely retarded oxymoron.
A proof is never empirical. A proof always on the side of speculations passed as rationality.

Logic is just a field by autistic pedants about well formed formulas and valuations, ie a map sending a formula to 1 or 0 and asking what are those valuations which are stable under inference rules. Zero truth in this, especially truth in the casual sense. Tarski truth is moronic, meaningless. Peak atheist. Just like there is no truth in science, just some stats and a stat convention for saying ‘’if p value is XXX then the result is """"""""""""""’true"""""""""""

[Anonymous No.55339577]

>>55339572


At best scientists can come up about some stats about some formal system (ie a model) like the spring or harmonic oscillator or the standard model. Like ‘’your material has such and wear and tear, and our backlog of such conditions lead to 60% of breaking in the next year, therefore your material may break or not within a year’’ That’s the pinnacle of the scientific claim and all their claims remain phrased as uncertainty.

[Anonymous No.55339599]

>>55339572
so what's the solution? unbiased experience without expectations and plans?

[Anonymous No.55339694]

>>55339599
That's never any solution

[Anonymous No.55339713]

>>55339572
>>55339577
Better than believing boats float because satan powers them.

[Anonymous No.55339778]

>>55339713
yes, but the craft of sailing predates science. his argument was for empiricism. you don't need science to navigate existence successfully.

[Anonymous No.55339865]

>>55339778
Imagine the insurance policies.

[Anonymous No.55339889]

>>55339572
take your meds

[Anonymous No.55340263]

>>55339559
For a midwit it will.

[Anonymous No.55340331]

>>55339865
This is where actuaries come in

[Anonymous No.55340355]

>>55338073
No, but it makes me think of Ben from into the clown verse.
Trading is something which is so simple a 80 IQ ape can do it better than a giga high IQ genius because intelligence is a negative trait for decision making.

When you are trying to use something super complex to justify your decisions what you are doing is trying to remove responsibility from your own decisions which will always end up badly in this context as it's possible to manipulate data.

[Anonymous No.55340369]

>>55340355
This is such a retarded take i don't even know how to respond. Being disciplined enough to keep your winnings and minimize risk is not simple. If it was, you'd see more low iq retards doing it or just getting hired by hedge funds to manager money. Something you never see because no one would ever trust a low iq person to do this. They have to filter them with math and engineering phds to even consider an interview.

[Anonymous No.55340664]

>>55340263
You incestial illiterate.
As if you shitstsin can read PhD books and start earning billions within minutes.
Meme

[Anonymous No.55340757]

>>55340664
It would take less time if you weren't a midwit.

[Anonymous No.55341074]

>>55339572
nigger

[Anonymous No.55341129]

>>55340369
>Being disciplined enough to keep your winnings and minimize risk is not simple.
yeah, but it has nothing to do with maths. This is what nerds hate to hear.

[Anonymous No.55341155]

>>55341074
You're a master of sarcastic wit.

[Anonymous No.55341164]

Here's a statistic. The average American woman is 171 lbs. 78 kg for my European friends

Do with this what you will

[Anonymous No.55341184]

>>55338073
>studied statistics
This, plus a minor in finance, will get you a Wall Street job paying millions.

[Anonymous No.55341211]

My life looks like "Gamma"

[Anonymous No.55341273]

>>55341129
Lol only a brainlet would say this. Getting a phd in math shows your employer that you have discipline you low iq faggot. If you cant even see that then you have no business managing even your own finances.

[Anonymous No.55341461]

>>55338073
Went to Stanford for stats
Unless you're in hft it's a joke
I currently use neural nets for trading though
Hard core stats in tradi g don't work outside of very short time frame approaches as randomness creeps in along with regime change from news pouring in
Neural net goes brrrr tho

[Anonymous No.55342333]

>>55341273
yeah weird how the all the nerds cant beat the market tho, fuckking retarded undergrad zoomer

and having a phd doesnt mean shit in terms of trading, not even about discipline.

[Anonymous No.55342356]

if anyone else is interested in self-studying stats hit the discord zzz_yyy#7457
i've picked up some interesting open courses and textbooks
a strong foundation in stats seems like a good starting point for developing later edge

>>55341461
well that's good. i'm decent at ml

>>55340355
this is denial

[Anonymous No.55342802]

>>55342333
Hedge funds beat the market retard. Go argue with your fellow zoomie undergrads about retail trading while the big boys play.

[Anonymous No.55343730]

>>55339572
based

[Anonymous No.55344693]

First informative thread on biz in a while I will not let it die. Post stat textbook recommendations

[Anonymous No.55344740]

I'm writing some scripts in Ninjatrader looking at the pairs spread difference between /ES and /NQ (or other pairs) and normalizing them. Looking for trades outside of Z > 2 or Z < -2. Still in progress though.

[Anonymous No.55344783]

>>55344740
interesting. i program as well.
discord? >>55342356

>>55344693
https://drive.google.com/file/d/1VmkAAGOYCTORq1wxSQqy255qLJjTNvBI/view
is a free one by the harvard STAT110 professors. its good

[Anonymous No.55344834]

>>55340369
>Being disciplined enough to keep your winnings and minimize risk is not simple
>this
imagine my surprise when i learned selling is the hard part

[Anonymous No.55344879]

>>55340355
>When you are trying to use something super complex to justify your decisions
its surprising how true this is. all the complex up your ass solutions suck and the simple ones work much better

[Anonymous No.55344883]

>>55338958
>>55344693
For complete beginners? Thinking With Data is a good book written by my old prof.
https://people.uleth.ca/~vokey/pdf/thinking.pdf

In grad school Designing Experiments and Analyzing Data by Maxwell & Delaney was my main reference manual. But I didn't really use textbooks often, online tutorials moreso. Learn R (use RStudio to make it easier) and you pick up plenty of stats while studying different methods.

[Anonymous No.55344890]

>>55338073
all the statistics autists with their mathematical probability distributions lead to the biggest risk exposure and that lead to crashes

[Anonymous No.55344910]

>>55341164
kek

[Anonymous No.55344915]

>>55344883
have you read probability: theory and examples by rick durrett?
seems like the pinnacle of rigorous stats

[Anonymous No.55344924]

>>55344783
I don't do discord, and this is still very much a work in progress. I'm a believer in market profile, but I could never figure out the best entry and exit times. Here's the basic idea though. This is from Friday. I usually start programming in ThinkorSwim because the scripts are way simpler, but Ninjatrader has better data and cheaper commissions.

[Anonymous No.55344927]

>>55341461
>I currently use neural nets for trading though
please elaborate Anon...

[Anonymous No.55344945]

>>55344915
Quit bumping your dead thread faggot fucking kys

[Anonymous No.55344955]

>>55344890
>>55339572
the market promotes ignorance in the public because they fear losing their edge

[Anonymous No.55344960]

>>55344924
looks interesting. are you collabing with anyone on this or just solo?
you should try discord, other anons are sometimes helpful

[Anonymous No.55344966]

>>55344955
they want the gentiles dumb and clapping at magic tricks and believing god forbids interest so they can maintain their corrupt empire

[Anonymous No.55344967]

>>55338073
All you need for successful trading is RSI, MACD, your eyes and a brain.

[Anonymous No.55344979]

>>55344960
Just solo. Here would have been the next few trades that day.

[Anonymous No.55344981]

>>55344966
throwing in their retirement into their ponzi schemes and believing the market cant be predicted. keeping the margins fat and non competitive

[Anonymous No.55344993]

>>55344979
bretty good if you can figure out those exits

[Anonymous No.55344996]

Kill yourself spammer
>thread on last page, spams it to bump
>hops on device to create a false conversation to keep his thread up
I really want to know what runs through the head of you faggots.

[Anonymous No.55345015]

>>55344915
No, but looking at the table of contents I'm familiar with most of the concepts. My approach was find something I'm interested in/need for a particular experiment (e.g., causal analysis, reinforcement learning, hierarchical bayesian models), then study the underlying concepts (e.g., advanced concepts behind the central limit theorem, random walks, Markov chains, Volterra decomposition, signal detection theory), so that I can properly do whatever it is I want. A lot more engaging that way when every concept has to do with a specific project I'm working on.

[Anonymous No.55345046]

>>55345015
>but I just DCA into Bitcoin regularly.
with all this stats knowledge how come you don't do any quant stuff? capital requirements are too high for effective strategies or what?

[Anonymous No.55345074]

>>55344993
I was also looking into intraday stats based on initial balance, inside bars, outside bars, etc. on ThinkorSwim. It's not something I'm putting a ton of time in since I still have a day job, but my goal is to use these as targets and write a script that shows high percentage prices based on recent price action. These are based on 30 min bars.

[Anonymous No.55345075]

>>55345046
Mainly because I don't care, which is funny because I studied economic decision-making using reinforcement learning models, but I honestly can't be bothered.

[Anonymous No.55345147]

>>55345074
did you get these target %s through backtesting? this is pretty interesting stuff. with the appropriate statistical models surely you would be able to tell if there's anything here statistically significant for an alpha edge

[Anonymous No.55345183]

>>55339572
I want to cum relentlessly is tight virginal pussy, this is all I want

[Anonymous No.55345191]

>>55345147
Answer me faggot why are you ignoring me and spamming the catalog.

[Anonymous No.55345196]

>>55345147
Writing scripts that count the number of instances. What's interesting is that these statistics are pretty consistent across liquid assets. Here's SPY vs AAPL. Each line represents a counted event. The cyan line is the total bars, so the percentages are out of the total number of bars.

[Anonymous No.55345210]

>>55345196
gaps must be created in areas where the majority's entry criteria get fulfilled

[Anonymous No.55345256]

>>55345210
Not sure what you mean by that, but disregard the rectangles showing the gaps. Those aren't related to the scripts. Each of the lines is a different event. So for SPY, there were 2328 bars, and whatever the logic was for the pink line happened 1429 times out of those 2328 (approximately 61% of the time).

[Anonymous No.55345263]

>>55345196
you should check out the market wizards series of books by Schwager. most of the traders there are using qualitative methods, but a few (esp. in the latest book) use somewhat similar quantitative methods as yours if i remember correctly. one of the guys was an algo trader making consistent double digit % returns every year for several years

[Anonymous No.55345316]

>>55345263
I'll check them out. Thanks for the recommendation.

What's more interesting though is that the statistics often hold up on different timeframes. This shouldn't be too surprising with a fractal market, but still cool to watch it play out.

Pink logic on 5 min = 64%
Pink logic on 1 min = 62%

[Anonymous No.55345318]

>>55338073
Of course. Here's my conclusion OP. Big green chart go up? Good thing
Here, i made a simple chart using my favorite memecoin (teehee)

[Anonymous No.55345322]

>>55345256
i get what you're doing with your scripts, your pic just made me realize that gaps are probably where the entry/exit criteria for a majority of traders is met and they all rush for the door at the same moment

[Anonymous No.55345337]

>>55345318
this is unironically true and highlights the problem with a purely quantitative approach

[Anonymous No.55345346]

[Anonymous No.55345387]

>>55339572
Of course science is useless as a method of gaining knowledge, because knowledge in the philosophical sense does not exist. All science (done properly) can do is make inferences about reality. The muh science people have completely lost sight of this, if they ever understood it in the first place. I honestly don't think most people who work in science really understand the purpose or limitations of science, it really is some sort of secular materialist cult for them.

[Anonymous No.55345526]

>>55338073
>has anyone here rigorously studied statistics and is now using it in trading?
yup, the options folks have the most in my experience

https://www.tastylive.com/shows/the-skinny-on-options-math

[Anonymous No.55347123]

>>55345526
primarily why i wanted to learn stats in the first place.
but also interested in applying for other quantitative methods

[Anonymous No.55347215]

>>55344927
was on mobile on a plane, same guy here
I basically make a neural net to shit out a number that determines if the price will go down or up in the future
how I do that, and then process that number to make a decision
thats the hard part

[Anonymous No.55347265]

>>55339572
Your last point, it’s not that “X is true”. The P-value is simply a probability of the null hypothesis being true, under the assumption of the alternative hypothesis being true. It’s basically a probability of some other random phenomenon explaining observed results.

So when you get a P-value of 10e-7, it’s a very small probability that the observed phenomenon can be explained by something else.

[Anonymous No.55347376]

>>55338073
Did stats and actuarial math. I don't think that shit works for trading, and best case it's just a rough estimate of what they trying to measure, and when actual happens, it is usually off from the models.

[Anonymous No.55347379]

>>55347215
what % of your trades are profit and what % are loss?

[Anonymous No.55347475]

>>55347265
Technically it's the probability of your data being due to random chance if the null hypothesis is true. The probability of the null hypothesis itself being true or false can be derived through Bayesian means however.

[Anonymous No.55348921]

>>55338073
I use Stats at work in modelling and process capability and studies for improvement. I found using Standard Deviation is reliable and found that idenitifes relative extremes to price. Example is using a Standard Deviation Channel in ToS. When a stock moons, it usually hits no more than 2 SD from mean.

[Anonymous No.55350949]

looks like i won

[Anonymous No.55350996]

>>55339572

[Anonymous No.55351165]

>>55338073
Any good references to get started with ML and applied stats/quant in a programming context?
I always get bogged down and loose patience due to minute details (often relating to implementation & tech details) so would be happy to know of references which do not forget to cover it.