/smg/ - Stock Market General

Please stop arousing me on the Lord's day

Greeeeeeeeen

Attached: basedgoldenbull.png (1056x592, 399K)

Its the editions fault not mine!

Attached: endcard02.jpg (800x451, 154K)

early for long netflix

Attached: netflix-logo.png (1280x720, 105K)

I'm so excited for tomorrow. I took work off so I can sip and watch the market all day.

Attached: Star Wars Episode III - Revenge of the Sith.(2005).1920x816.x265.2ch.AAC.mkv_snapshot_01.14.52.jpg (1920x816, 168K)

I'm fully prepared to be raped, but on the same hand, I am prepared to rape.

Attached: Kiss-shot.Acerola-orion.Heart-under-blade.600.1965259.jpg (518x600, 40K)

Threadly reminder that all tripfags should be required to do the their all time P/L.

Attached: 1531674228111.jpg (750x541, 194K)

hope this iraq stuff makes my 7-20 USO 14 calls 3 baggers.

Attached: 1474389119469.jpg (850x599, 134K)

what stocks will you be watching? better to take of tuesdays-thursdays imo.. mondays and fridays are kinda boring in the market.

If SPY doesn't crash into the ground, please rape me to death

very cute

give it time. we need to see how the first couple ERs go. if its like last time, where 75% of the record ER beats result in 5-10% drops, then i have no doub that were going down in the near-term.

watch the SMI (smart money index). its headed STR8 down. and it hasnt slowed down at all. smart money has been in "sell the rally" mode since end of jan. if last quarters ER beats werent enough to convince them, and we see the same trend for this quarters ER beats, then its pretty clear where were headed in the short to medium term.

if were truly fugged, then they have to keep it propped up until they get all their shekels out. that means suckering retail into buying their bags BEFORE pulling the plug

>smart money index
they should really change that misleading name considering it has nothing to do with large specs

Short Tesla.

I'll for sure check out the SMI, sounds legit. My puts expire on the 23rd, so I need to get out on the 17th-18th to avoid getting stuck with them.

Ok, sick graph.

>thinkingface.png
Ok, I'll take that.

Why? More trade war shit?
I don't recall any new information in the news

What Iraq stuff?

What would a better name for it be?

So I was in on ALT at .43, where’s the exit on this one? Is there long term potential to hit $1+ or was Friday the pump?

Thanks frens

just a man doing his best, how ruined am I?

Attached: stock 7-15-18.png (1531x747, 113K)

It's diluted because I've deposited a lot since starting, but still beating the indices

Attached: Screenshot_20180715-154409.png (1080x1920, 108K)

>green on GE
well done user

scoops, I read that whitepaper. interesting stuff.
things I like
>nonlinear over linear models
this is essentially reducing the rationality in the models by drawing meme curves instead of lines
>careful attention to avoiding forward looking bias
>attention to meta parameters
>frequent retains
though this is a double edged sword
>breaking up by sector
a decent proxy for controlling for macro

things I don't like
>length of tail for training data
>meta parameter choices
they essentially hand tuned these values to maximize their results. also these were both static. much better if they were adaptive. I particularly also don't like 3 years tail for quarterly positions. I think it should be a longer tail but then that's based only on instinct
>no proper incorporation of macro changes
though to be fair this is really hard, but I bet weighting in a couple of indexes could do it
>training set 1983 to 2010
there's a lot of past bias here, and I worry that it doesn't account for the modern explosion of algo trading

Attached: 702bc4ff-c484-4347-8fd3-9aac83655992..jpg (900x686, 87K)

the nobody index

"index tracking people who sleep in"

High chance that AMZN gets their boxes from SON or RKG anybody have any other companies that might be supplying AMZN with boxes?

oops i meant PKG

th-thank you user

Its earnings season, no matter what you hold or trade massive PR waves come out about everything. Whether your a pajeet, chink, or goy, shit will happen. In my case, Netflix will effect IQ.

Kek. What's your favorite technical indicator?

TA is a crutch
You only need to glance once at the price and you know

wsj.com/articles/iraq-struggles-to-contain-protests-over-government-dysfunction-1531675702

>scoops, I read that whitepaper. interesting stuff.
I still need to just repost the whole link bundle :/

>things I like
>nonlinear over linear models
>this is essentially reducing the rationality in the models by drawing meme curves instead of lines
Lots of shit just doesn't fit well with linear models
This includes a good explanation and isn't very long:
datacamp.com/community/tutorials/svm-classification-scikit-learn-python
Using something "too nonlinear" results in overfitting, and RBFs are generally a pretty good middle

>careful attention to avoiding forward looking bias
This is a must. Otherwise, it's training off the answer key
>attention to meta parameters
Just using a grid search of C and gamma is all this really means imo
>frequent retains
>though this is a double edged sword
Yeah, I'm not planning to do that. The recency of my data shouldn't matter for what I have planned.

>breaking up by sector
>a decent proxy for controlling for macro
Yeah, I liked this. But, I'd prefer it to be a parameter on its own instead of completely separating the models. However, backtesting should be used to see if combining the models and just using a parameter will degrade prediction accuracy

>things I don't like
>length of tail for training data
>meta parameter choices
>they essentially hand tuned these values to maximize their results.
Correct. But, backtesting is supposed to prevent this parameter tuning from resulting in overfitting
>also these were both static. much better if they were adaptive.
The parameters can't be adaptive -- that'd be like switching back and forth between a linear and polynomial model for whenever it's convenient and gives better results
>I particularly also don't like 3 years tail for quarterly positions. I think it should be a longer tail but then that's based only on instinct
I'm mainly just stealing their TA and FA features and planning to experimentally determine what timeframes give the best predictions.

>that boomer looks at the price

buy red, sell green. Ez money

>no proper incorporation of macro changes
I do plan to throw in some shit like the yield curve and momentum of TLT's prices
>though to be fair this is really hard, but I bet weighting in a couple of indexes could do it
I'm going to at least throw in some features that follow SPY, but DJIA is based off 30 hand picked companies and NASDAQ is like 95% the same so those probably won't go in.
>training set 1983 to 2010
>there's a lot of past bias here, and I worry that it doesn't account for the modern explosion of algo trading
I like having lots of history. That's more data. I want to at least include a couple recessions of data in there.
Also trading will only be an issue if the model does daytrading, which isn't the timeframe that I have in mind.

its stil a massive amount of wealth sneaking out the backdoor at the same time. and it hasnt done anything except go straight down since february, so i can only assume that the event thats being sold off en masse for, hasnt been realized yet.

Elon is so fucking bad at twitter.

"Generally the view that I've had on Twitter is if you're on Twitter, you're in like the meme—you're in meme war land. If you're on Twitter, you're in the arena. And so essentially if you attack me, it is therefore OK for me to attack back," he said.

First for Bitcoin vix indicator (not pretty)

i saw this meme posted here the other day. ngl it gets me shakin in my boots

Attached: 1531552079668.jpg (890x468, 52K)

Is there a way to see a list of stocks trading below NAV?

Elon is utilizing the Trump effect, attention will only make him more powerful

What?

you see, a lot of money has been made in the market so it only looks bigger than the reality. Its just a drop in the bucket these days, and large specs know the truth, and arent selling

Attached: xzc.png (361x411, 260K)

>GE
I guess now might be the time to go long.
I shorted the shit out of them until the mid twelves.

nah, he doesn't have the knack for it. Trump does it in a very unique semi-playful way.

After reading up on what it actually tracks, I think it just shows that you can currently buy at 3:01pm and sell at 3:59pm and make a nice profit...

this type of idea getting very popular, you aware of Super Guppy?
not bad for first script

Finished making this one,
this is all the SPDR sector funds on screen + spread of all XL averaged.

idea is SPX and XL Sector Funds averaged are basically the same thing which we can see kinda true

Attached: XLindex.png (1804x857, 206K)

tradingview.com/scripts/guppy/

Guppy!

on one hand it makes sense that the sell off in the SMI is actually larger, as were UP higher, and all the stocks are worth more per share ( hivemind). but its literally only gone in one direction since volpocolypse. i think at this point, its the largest divergence ever in history of it being tracked. and its definitely the longest its gone in a divergence w out the markets crashing

>you can currently buy at 3:01pm and sell at 3:59pm and make a nice profit...
now think about who trades on bells and think about why that would be happening...

people are happy with the overall gains up till now. It's just very defensive play and not necessarily anything else. 2007-2008 spooked everyone and it's still on minds to avoid the next one. It's extremely predictable behavior.

only 1.5 hours left until Futures gap up and liquidate you

>who trades on bells
every trader
Small specs are thinking shorter term
aka bullrun golden imminent while weakhands got rekt
baaaaased

Attached: 104258723-GettyImages-104554645.1910x1000.jpg (1910x1000, 175K)

>Using something "too nonlinear" results in overfitting, and RBFs are generally a pretty good middle
>Just using a grid search of C and gamma is all this really means imo
You're thinking far too narrowly. markets are irrational and I'm confident that an ml strategy which better approximates irrationality would produce significantly better performance than what's essentially hyper rigid ta. I've had great success doing cc fraud detection and weather pattern prediction with adaptive, self mutating graphs
>The recency of my data shouldn't matter for what I have planned.
enjoy getting your pants pulled down by major events or even earnings reports. the whole point of ml systems is adaptive behavior based on new incoming data
> I'd prefer it to be a parameter on its own instead of completely separating the models
this will introduce some significant amount of regression to the general market mean.
>backtesting is supposed to prevent this parameter tuning from resulting in overfitting
back testing is how they did the tuning. they've used it to fit to the training data, introducing significant selection bias
>The parameters can't be adaptive -- that'd be like switching back and forth between a linear and polynomial model
no it's adjusting coefficients and weighting, it doesn't change the the model. they even state outright that not adapting the meta parameters is a weakness

Don't say such terrible things, user
Reminder the CHAD Nasdaq beats the VIRGIN dow

Also keep in mind the market is effected by technology increases. As we see the normal up and down cycles keep in mind that technology is improving. It was largely wasted on the focus for advertising but it seems to be increasingly moving towards other areas such as the physical world. This can be most seen with the emphasis on self-driving cars from Waymo and Apple.

The entire crux of the tech valuations and entire market right now is literally entirely hinged along self-driving cars. Which are a proxy for machine learning's ability to influence the real world.

Basically, try to get as good as a predictive ability on how well self-driving cars will work out in the next 12 months. If you are bullish on progress then you should have all of your money long in relevant tech companies. As such success will overpower all recession possibility.

dow looks as solid as it comes. summer consolidating still in play tho

Attached: 43287946798de5d14fdfa7f1e5593ca3.gif (460x510, 2.93M)

>I do plan to throw in some shit like the yield curve and momentum of TLT's prices
>I'm going to at least throw in some features that follow SPY, but DJIA is based off 30 hand picked companies and NASDAQ is like 95% the same so those probably won't go in.
I had rkg's derived penis index in mind
>I like having lots of history. That's more data. I want to at least include a couple recessions of data in there.
>trading will only be an issue if the model does daytrading, which isn't the timeframe that I have in mind.
missing the point entirely and wrong. I said nothing about your model doing day treading. algo trading in general significantly impacts the way the market functions these days and historical focused data will train off outdated signals. including large corrections is important though

the DOW is more fierce

How well do ML algos work? Seems almost pointless to try ML unless it's just to refine already existing algo trading.

I don't really see the point unless you are doing market-wide and world-wide market super complex correlation type stuff. You would have to input so much data to get results it seems silly for a retail to try it, unless you input your own sentiments to get some help on planned trades.

/Bull_Dogs round up/
We have Cisco on discount
We have CAT on discount
We have PG on discount
We have BAC on discount

Pump Pump Pump.

Attached: 1523236146853.jpg (250x237, 23K)

You guys talk about technology affecting companies and industries, but really it's affecting the nature of trading itself

which is why I don't bother with any numerical related research. They are better than me.

Is trading on the /RKG/ + Guppy the ultimate way to success??

I CAN'T WAIT TILL 4 FOR FUTURES, I NEED FUTURES FUTURES

Attached: 1516567909381.png (550x445, 9K)

why, they are going to be green, dow +120

I haven't done any work with the stock market. You need to have good domain knowledge to craft quality algos, plus understanding of ml to pick the right strategy. It's essentially leveraged ta (doing ta on ta) in this case, with the added benefit of being able to act faster on action than a human could. But flaws in your ta or meta ta get magnified quickly

>algo trading in general significantly impacts the way the market functions these days and historical focused data will train off outdated signals. including large corrections is important though
i agree. whenever i sit down and start studying historical data, it always strikes me how different the market behaved on a day to day back then, compared to now. algos definitely changed the way things move.

dunno how this would effect ML tho. id assume, the more data, the better

waymo jaguar

waymo plans for 1 million taxi trips by summer 2019.

20,000 planned electric vehicles from jaguar, built to have self-driving support.
82,000 ordered from Fiat Chrysler

Attached: 7bfae7f5-fed4-4e4c-a73c-cd87a2e18a83.png (670x357, 354K)

This. Musk going to get btfo Monday by his leddit "investors" over rnc contributions plus calling dude a pedo on twatter

yeah down early in the day, recovery by afternoon seems most likely.

why are you all buying into an already overbought market?

where my /shortsellers/ at?

Why is Nikkei doing so well?

Attached: 1531666523883.png (600x800, 629K)

>The more data the better
Not necessarily. Overfitting is a primary concern when making models, and it's using too much data, making your algo susceptible to noise or random blips. You have to carefully discard a % of the input data to get best results

Use handcrafted abstract syntax trees to create bottlenecks?

its not...

T-those can't be real

>bought tvix at $38
>it's earnings week
Time for tvix to drop down to 30$ this week baby!

You can see distorted pixels around the boobs, don't be tempted by 3DP.

Attached: 1530019660483.jpg (750x1061, 374K)

I guess you could use asts if you're using a semantic model. But almost always it's quantitative, so fitness functions and parameter tuning. Even the graph traversal work I've done used quant decision functions and wasn't pure ast style semantic choices
Bitches don't know about my hcup jav

Loaded up on sqqq at close Friday. Hoping for some pic related

Attached: 1529703813365.jpg (720x881, 170K)

saki yanase is real, stupid animenigger

T_T
pretty much everyone thinks we about to hit the ranging supply zone 2800-2840 that will fuel the rocket moon to 3000

gonna add some buy / sell signals to that script to see how effective it really is
need that ishraes boxx guy to identify more patterns @_@

Attached: 61Rb+NheXwL._SL1500_.jpg (1500x844, 75K)

Post butthole

Kill yourself

Kill yourself

Attached: 4321432115.jpg (459x721, 135K)

30 more minutes
I can feel the green day already
woooooo

Attached: 1530673430183.jpg (755x687, 189K)

Hows your butthole feeling by the way?

>analize
Subtle

>wait for futures
>see futures are all green
>be bored again till markets open
>watch china and asia go green
>wait for USA open
>tesla down 5% and market green

this loop is boring

Like it about to be gapped open by gigantic throbbing red hot glicining ES_F

that was on purpose and totally no byproduct of mspaint not have spell checker... yes

Attached: 20180628_120214.jpg (430x277, 73K)