Hacker News new | ask | show | jobs
by stockstream 3265 days ago
Hey everyone! Glad to see you're all enjoying the project. I'd be glad to answer any questions you have. Let me know!
11 comments

You should do a similar experiment with BTC, or BTC+litecoin+eth or something like that. The APIs are already there.
That wouldn't work.

The only reason he is getting away with doing this for stocks is that Robin hood doesn't charge any per trade transaction costs.

Day trading cryto would kill him in trading fees.

Trade on a platform with negative maker fees with post only enabled.
he has 50k on the account, he can pay the platform fees. if the returns minus fees are higher than stock returns sans fees, it is game. and right now, it might as well be.
Bitcoin required a couple dollars per transaction, not to long ago.

You are overestimating his risk, right now, for stocks.

The current setup isn't likely to lose him much money. Throwing darts to pick your stocks will cause your stock portfolio to be surprisingly diversified.

Random stock pick is actually a not crazy investment strategy that is likely to get you the same amount of return as putting your money in an index fund.

Transaction fees in cryto, on the other hand, would cause him to lose 10k a month.

Trades on an exchange aren't on-chain, so transaction fees don't apply. The fees are whatever the exchange sets.
And those fees are usually quite high. Like, around 0.2% to 0.5%.

This is actually HIGHER than the on chain transaction costs.

Yeah did some initial research, but didn't find a good platform with an API and no fees.
You can place limit orders with no fee on GDAX/Coinbase, might not be quite what's needed though (I have zero knowledge in the area).

The main downside is that as soon as Bitcoin goes super volatile Coinbase will go down like clockwork.

Bitmex has an excellent API. 0.75% fees but -0.25% if you place limit orders (with post-only on)
Any plans to weight more highly the advice of those who have previously given good advice?
Thought a lot about this during the development period, but nothing materialized. I might do it someday but would probably need to ramp up the scoring algorithm first.
Why did you decide to do it with $50,000 and Robinhood instead of just creating a bot to trade stocks on an Investopedia stock simulator account? [0]

I mean, I understand that $50k might not be a lot for you, but if the aim is to see what the majority of random people on the internet feel like buying, while measuring the return of these decisions, this could be accomplished without putting up any actual money.

[0] http://www.investopedia.com/simulator/

One of the reasons I decided to do this project in the first place is because nobody has ever done this before (at least based on my research), and I wanted to be the first. There are plenty of stock simulators out there, so fake money didn't sound appealing to me.
The actual money aspect is why this blew up. No one would care if it was fake money.
> The actual money aspect is why this blew up. No one would care if it was fake money.

Two things:

1. This is assuming it was done for the press coverage, and not just as a hobby project

2. If I were to do this, I don't think I'd advertise it was fake money. What's the difference anyway? It's not as if people on the internet benefit at all from the investment return. I always assumed this was more of a "Twitch plays Pokemon" thing than anything more serious.

I wasn't assuming it was about press coverage necessarily (though maybe it was), but that in order to be interesting it needs a critical mass of people participating. If it wasn't real money then that wouldn't happen.
Random people will make different choices because it's real money, for one.
Any plans to open source the software you're using to glue Twitch to Robinhood's APIs?
Actually, those two bits are pretty small. Theres a ton of code all over github for interacting with Robinhood. The twitch chat API is really just IRC. Glueing those together was the easy part.

The more interesting and complicated components are the UI and the scheduling and a publish-subscribe module that I created, tho no plans to open source it anytime soon.

>publish-subscribe module

What part of application is that for ?

Actaully the pubsub part is pretty core to system overall. Basically I have a set of 'operations' modules each operating independently (gathering news, tweets, votes, stock quotes, etc). They publish that data over the pubsub system, so that all the subscribers (mostly in-memory caches) read and update accordingly.

Game events also get published through it (round end, trade placed, new round, etc)

Hi Mike!

Was talking with your folks about your project at various recent Langley weddings.

Kudos for a great idea, well-executed!

Thanks! Seems word spread through the extended family pretty quickly.

We should stay in touch though, add me on LinkedIn if you use it:

https://www.linkedin.com/in/Mike--Roberts/

I am just wondering how you would do your taxes next year.
Just gonna hire someone to do it for me.
Shouldn't really be any harder? 50k going in, pay taxes on profits, or get deductions on losses based on where things stand December 31st (that is, assuming there are any taxes at all before liquidation) ?

(And of course pay taxes on any dividends)?

You need to report each trade to the IRS, plus some stocks are pass through in term of tax. Mainly oil stocks. It means you have to report their business expenses, profits, and revenues inside your own tax returns even with only 1 share. It's not a big deal, but that's result in lot of paperwork.
Interesting. What's the rationale for this? Detecting market manipulation?
It's just how you do the accounting. For instance, you might buy a stock over time, and then sell it over time as the price is rising. You have to pay tax on those gains, but how you do it is a choice. You can do first-in-first-out, so the gains would be e.g. on the first 5 shares you bought vs. the first 5 you sold. You can also do first-in-last-out, which might make sense if you can get long-term capital gains rates on some trades.

It's kind of ridiculous, but most traders will have some sort of automated solution to help with it. The exchange you use should give you a nice report of your trades.

Wait, wait, don't I know you?

You worked in AFT Picking, right?

Ah yeah, hey, how's it going? I remember you from the YYZ office.
Currently working in Fintech, so my coworkers and I had talked about stockstream before. Had no idea it was you until I opened the link above and did a double take.

Coming back to Amazon Fulfillment soon. Just couldn't stay away.

How much money do you actually have?
Heh, well I definitely don't consider myself to be super wealthy. $50k is not a small amount for me, but not the end of the world if I lose half.
The way your stream works I can't imagine you would lose that much, since the buys are in such small lots. There'd have to be a way for the users to make you go all in on a volatile stock to see those kinds of losses (or access to a sharper knife like options).

Anyway, it's an interesting experiment. Very popular over at /r/wallstreetbets.

Was looking through the comments to find a fellow yachting club member. Avast Matey!
Can I throw in $10K of my own to be invested, but take out the profits later?
If you'd really like to, just scrape the twitch chat yourself, count the votes and invest accordingly!
Haha nice. I wish I had the time....
Serious Q: how do you personally deal with the trolls?
There are no protections against trolls. At this point a dedicated group of less than a hundred could probably take over the vote.

I'd actually love to see more dedicated troll attacks. So far the main attack seems to buy $SEB since it's so expensive, but sometimes it later gets sold at a profit.

Personally, I've mentally written off the $50k so, I'll just be enjoying the show.

So in other words, you're basically counting down till the point someone posts it on 4chan or another notorious image board somewhere.

Good to know you've written of the $50K, because if those guys decide to mess around with this, that money is basically going straight out the window.

But how do you deliberately lose money? I think there's this idea that you could tank a portfolio with ease, but if that were true, shorting would be incredibly lucrative.

It's just as hard to lose money on stocks as it is to gain them (aside from racking up trade fees or something).

Isn't the ability to buy stocks that are going to go down effectively the ability to (short) the same stocks and make infinite money?
How do you think this experiment will play out?
No idea. Was bearish at first, but lately a lot more bullish. Only time will tell.