The Trading Game | Hacker News Reader

The Trading Game

bloomberg.com

235 pointsscottkduncan99 comments

The game definitely leaks information, I got better at every try -- this was the 4th attempt:

Your returns were 329.0%, against the stock’s 139.7% and the S&P’s 14.7%. You outperformed 100.0% of players, ranking #1.

Yea, the re-scaling algorithm is a future leak. I wonder how the developers didn't notice.

Did notice, didn't know what to do about it. I think it's the main problem with the game. There are plenty of options that don't leak and maybe I was just too enamored of the visceral feel of the y-domain following simplified+smoothed price line. Scale is also sensitive to variance over time, also smoothed, also leaks.

-the developer

Yeah, there is a bit of a leak but relying on that will not benefit you when you get Enron.

+1 for Enron. I was getting really excited as I saw it scale out and out and out. I knew there was going to be a huge jump and I was going to make a massive return. Unfortunately it didn't occur to me to be more reflexive. I think I barely managed to recoup my initial investment.

> Your returns were 5257.4%, against the stock’s -75.3% and the S&P’s 57.1%. You outperformed 99.2% of players, ranking #14.

interesting, but ulitmately, just a game about fast reflexes...

It's not. stock wiggles are very close brownian noise with an average direction of up at ~6% a year (random walk, EMH, etc)... and yet, playing this game, I seem to be able to time the tops and bottoms of my trades. I think the camera is leaking directional information in how it pans.

"I think the camera is leaking directional information in how it pans."

Which would make it a game about fast reflexes.

Wow. Imagine your livelihood depends on that game; your salary, your bonus, your standing. You can research the movements and HFT the micro-movements, but you're still playing that game.

And then imagine the same thing but you're not playing the game. That's what happens when you have massive exposure (ie work at a company) without doing anything to manage it. Diversification is good.

That's not really true. If the company goes bust, it doesn't affect your net worth at all. All that happens is one revenue stream dies, and you are free to replace that revenue stream. Your total portfolio doesn't go down at all. For most changes in share price you are completely insulated.

It is true if you are in the unfortunate position of havig a good chunk of your pension tied up in company stock. This is a terrible idea for the pensioner, of course, but nevertheless many 401k plans and a good chunk of defined-benefit plans put substantial exposure into the employing company's stock or debt.

They're becoming less popular after the slew of "company goes bankrupt -> retired workers lose their income" stories that have happened in the past few decades.

From what I understand about HFT, if you know what you're doing you basically can't lose.

To understand how, you need to understand 'put options', which is a financial product available on the derivatives market. A put option is a deal to sell company stock at a pre-agreed price. Worth noting that you don't need company stock before the put option deal is reached.

So let's say the stock price of Company X is currently valued at $50, and you arrange a put option to sell stock at $50. If the stock price goes down, let's say to $25, you buy stock at $25 and sell it at $50, so you've made a profit (from what I understand, you don't even need to buy the stock in this case, the put option broker will just pay you the difference to simplify the process).

But what if the stock price rises? Let's say the stock price rises to $75. You can't buy stock and use your put option without losing money, and from what I understand you pay interest to the put option broker for the length of time you have it, so holding onto the put option causes you to lose money.

But there's a 'get out of jail free' option. When you arrange the put option you also buy stock. If the stock price goes down, you buy more stock and sell at the pre-agreed put option price. If the stock price goes up, the stock you hold is worth more and you can use this to clear the put option without losing money.

From what I understand, this is one example of 'hedging' against stock price changes, there are probably others. HFT is well suited to this kind of deal, as you can react quickly to market changes, minimising any risks resulting from delays.

So although people see the stock market as gambling, you can game the system to move the odds in your favour. The end result is huge volumes of money invested into non-productive uses of money (and because of the ways banks create money, and the ways the financial markets are regulated, the restrictions on this speculation is basically non-existent).

EDIT: I've got a downvote on this already. If anything I've said is untrue, then call me out on it.

Who do you think is selling you the put option? Do you suppose they price those such that a trivial stock pair hedge will make money? If they did that they would be losing a ton of money & you would have replaced equity risk with counterparty risk as they go out of business.

HFT are good at options trading for the same reason all options traders have been throughout history. They can find misspricings in the market and trade them while they last. HFT is taking over those trades because HFT is much cheaper than the humans they replace & can therefore take on less high margin mispricings than humans can.

Also options trading is a regulated marketplace much like equities & commodities trading.

Finally 'yummyfajitas is an ex-HFT who has written good blog posts on the subject.

>"Do you suppose they price those such that a trivial stock pair hedge will make money?"

The example was just for illustration purposes, a simplified version of how it works.

>"Also options trading is a regulated marketplace"

Is that right?

https://en.wikipedia.org/wiki/Derivative_(finance)#OTC_and_e...

"Over-the-counter (OTC) derivatives are contracts that are traded (and privately negotiated) directly between two parties, without going through an exchange or other intermediary. Products such as swaps, forward rate agreements, exotic options – and other exotic derivatives – are almost always traded in this way. The OTC derivative market is the largest market for derivatives, and is largely unregulated with respect to disclosure of information between the parties, since the OTC market is made up of banks and other highly sophisticated parties, such as hedge funds. Reporting of OTC amounts is difficult because trades can occur in private, without activity being visible on any exchange."

>"Finally 'yummyfajitas is an ex-HFT who has written good blog posts on the subject."

I would welcome yummyfajitas 's feedback. I'm not attacking yummyfajitas, if that's what you're implying.

> The example was just for illustration purposes, a simplified version of how it works.

Sure, but real options trading is much, much more complicated than that. You might as well have said "Find an equity that you know is underpriced and buy that". There is no secret sauce in options trading for HFT or otherwise. Like any market based trading profession you find a place you think is incorrectly priced, you take a position, and then you either make or lose money. The more often you are right the better a trader you are. There is literally nothing special about options trading vs equities trading in this regard (except maybe leverage as options can be dramatically leveraged).

> Over-the-counter (OTC) derivatives are contracts that are traded (and privately negotiated) directly between two parties

You could do private options I suppose (but HFT would have a distinct disadvantage here), but most options are traded on exchanges (I believe CBOE is still the biggest options exchange in the US) just like any other exchange traded instrument. The SEC, FINRA and CFTC all have rules about options trading in the US.

> I'm not attacking yummyfajitas, if that's what you're implying.

I didn't mean to imply you were attacking yummyfajitas, only that he probably knows about options pairs trades given that they are a very common trading pattern for anyone who has worked in the trading industry, HFT or otherwise.

You can lose because the option price takes into account how much the market expects the stock to move (the implied volatility). Your proposed strategy is going long volatility and delta-hedging to maturity.

Others have already chimed in, but I'll use a real life example. At the close yesterday, NFLX was trading around $100, you could buy a put option to sell with strike price $100 expiring December 18th 2015 at a price of around $7.00 a share.

Using your strategy, you would buy one contract and simultaneously buy 50 shares. Each contract comes in multiples of 100, so your put option covers 100 shares and you own 50.

If you didn't manage the trade at all until December 18th, you'll make money if NFLX is below $86 or above $114. You'll have lost money if it is anywhere between the two numbers, because the actual realized volatility of the stock did not live up to the implied volatility price (and for many reasons that is usually the case).

Even if you did manage the trade throughout, there is no easy free lunch. You could set a rule to close the trade the moment NFLX crosses over $120 or under $80, but then you'd be giving up the upside that it hits $150 or $50.

I'm a beginner to this field, but using a real world example sounds like a good way for me to learn, so thank you for putting it together leelin.

I'm a bit confused by some of the terms used (such as implied volatility price, can you explain what this means?), but assuming I've followed enough to understand... If the put option covers 100 shares, why would you only buy 50? Wouldn't it make sense to buy 100 shares?

You buy 50 shares, because in your proposed strategy you want to make money whether the stock goes up or down. Because the option strike price and underlying price are the same, the appropriate ratio of underlying stock to options is about 50% if you want to be neutral to the future direction of the stock movement (delta-neutral as they say).

If you bought 100 shares with 100 covered, then you are purely betting the stock is going up by more than the option price. You would break even at $107, but lose on anything lower.

Implied volatility is the amount the stock is expected to move as implied by the price the market is charging. Sometimes it is easier to think of an option price in terms of volatility rather than raw dollars, especially when you are making trades of the type you proposed.

In the Netflix case, the current price of the option "implies" that the stock will move plus or minus 3% per trading day; if it moves more, say 5% a day, you are likely to make money.

Of course, thinking of options in this way also means you are on board with a ton of assumptions in modern options pricing theory, and lots of smart people point out flaws and objections.

Okay, I understand about the 50 shares now, and the rest of your descriptions of how this work makes sense too, thank you for explaining this clearly.

Going back to NFLX example, I have one more question. You said that NFLX was trading at $100 yesterday, so where does $7.00 a share come into play? Are the shares not $100 a piece?

$7 is the option price. The interest in your earlier example.

Ah, thanks kasey_junk.

Okay, so I thought I knew enough to work the NFLX example, but I still came up with some mistakes. Unexpected values shown in blue here:

http://oi61.tinypic.com/2h71ndh.jpg

I've probably made a fundamental error, what am I missing?

Don't execute on out of the money options. IE you eat the $700 in fees but you don't actually trade the underlying stocks they represent.

Oh, so you can just pay the fees, cancel the put option, and keep the stock you bought at the beginning of the process?

Or really sell the stock you bought at the beginning of the process if you want to take profit.

Yes, that's one option.

So just so I understand, what flexibility do you have as the purchaser of the put option when it comes to selling. From what I understand the purchaser can choose to sell between the put option maturity date and the put option expiry date, is that correct? Are these typically the same date? Would you pay a premium to have a wider gap between the maturity date and expiration date?

The major difference between OTC and exchange traded derivatives is that exchange traded ones have set contracts that standardize these things. What those set contracts are is exchange to exchange and product offering to product offering.

You can take a look at CBOE's product web page to get an idea about the different kinds of options products they offer.

Taking into account put call parity, you've claimed you can beat the market by buying call options. Since you buy the call in one transaction you avoid all risks from delays regardless of being an HFT. Best of luck with that, let me know when you become rich.

Some HFTs do stat arb on put call parity, though I would hardly call such a strategy "can't lose".

The bit which is likely untrue is "you can game the system to move the odds in your favour." This bit is entirely unsupported by the rest of your comment.

Nothing you said is untrue; you just made it feel like math is some magic pot of gold.

Nearly every time, the cost of the stock plus the cost of the put option will be exactly the same as the values of all expected outcomes.

> "Nearly every time, the cost of the stock plus the cost of the put option will be exactly the same as the values of all expected outcomes."

Can you explain this further? Are you suggesting that you're obtaining the put option and the stock from the same source?

Yes. You would be purchasing both stock and the option from "the market."

As much as I like to joke about retiring to some farmers market to handcraft artisanal equity options, when people are talking about options on stocks (equity options) they are normally referring to exchange traded deriviatives.

Sure you can get industrial futures contracts, but don't you think your kids health deserves my free range, certified organic front month crude?

> "Yes. You would be purchasing both stock and the option from "the market.""

Does "the market" include OTC derivatives, i.e. options not traded on an exchange?

Depends on the context. But in order to get OTC options you are going to need to be doing some sort of massive deal (AFAIK never done OTC derivatives trading) or it needs to be some sort of esoteric option contract. You wouldn't do this as part of a normal equity/option hedge or trade. You also would almost certainly be paying a higher price for the privilege than you would on an exchange.

Buying a "normal" option contract on an equity is almost certainly not an OTC deal and is instead exchange traded. Those exchanges are very efficient like the equities exchanges are so options are priced competitively (thanks to options traders).

It's very easy to lose. Flash crash, transaction fees, your underlying model assumptions suddenly failing. Maybe you simply are doing worse than the S&P which is a failure in itself.

This game is really easy, so it can't be realistic. I think that a lot of the challenge is removed by the fast speed that the game makes the years go by. It removes the challenge of waiting with what you have for a better opportunity; I'm not sure if institutional investors would ever hold positions for years on end.

I'm not sure if institutional investors would ever hold positions for years on end.

This is extraordinarily common, particularly for institutional investors who are not attempting to generate alpha.

I think the main point being made by the author is that without the waiting, day after day, combined with the external confirmation bias of hearing how other stocks are outperforming yours (you thought you did great on your trade, until you hear Google could have made you multiples more); you are missing a ton of the statistical noise that screws up your trading.

Don't throw out the entire argument with nitpicking on a minor misunderstanding re: institutional trading. The main thrust of his argument is clearly valid.

If you try ThinkorSwim from TDAmeritrade, you can use paper money - I find the mini S&P to be a great start. Day traders are nuts though.

It's a game, no real money involved. Could try finding some source on this, but the challenge can only be measured when you're not using play money.

It's a shame I don't get 2000% returns in the real market. But I also don't invest 100% of my portfolio on every trade. I guess that's where I've been going wrong?..

The game should really be named 'The Momentum Trading Game'.

Without even knowing the company's price-to-book and p/e ratios, let alone what the company is, its past history, the market it's in, and who runs it, it's just chasing the dance of a number.

Doing nothing, I tend to beat about 20% of the traders. Without the future leak, it'd surely be better than that.

Edit: Just got Enron. Did nothing. Beat half the players.

The shifting of the scale on the axes gives you insider information...got to $1B after a few tries :)

Insider information is known but non-public. I think it's more appropriate to describe this as forward-looking. I.e. we are getting an insight into things that haven't even happened yet!

Yeah, their movement smoothing filter seems to be not causal.

I like "insider information" as a lazy excuse for the hinting/info-leakage problem. :) It's a metaphor!!

-the developer

I struggle to see if the game includes all the taxes and the cost of transactions? this is somewhat important on small gains

I prefer this version of the game. It's set to historical stocks from the early 1990s to today.

http://chartgame.com/

But it's made with flash :(

stocks from the 90s delivered with technology from the 90s!

Fun game, but real life trading is probably not that easy!

"Your returns were 2950.1%, against the stock’s -75.3% and the S&P’s 57.1%. You outperformed 97.1% of players, ranking #48."

I did ok in this game, i tried one of those plus500 demo accounts before and got slaughtered.

Information leak #1: Scale Changes - noticed this but could not put my finger on defining it clearly.

#2: Date - along the bottom edge the date progresses. Knowing generally that the market did well during certain multi-year ranges was an absolute edge. Too bad the real world doesn't come with this kind of information.

Hi I made the game. Really glad you brought up #1 — I think it's the main problem with the game. The y-domain follows a simplified and smoothed version of the price line, which, yes, leaks information about the future. Not sure what to do about it. Because viscerally I like the movement.

#2, yes, that too. We talked about that issue but maybe not enough. I wonder if you even need to know what the timescale is. Stock movements are pretty fractal and scale-invariant, after all, right? Kinda? Hm.

I like the date information as it gives a very much needed perspective! But, knowing the year itself may not be.

Could you drop the year indicator?

Pretty cool! Not something I thought would work so snappy on a mobile browser. I made a killing on Enron!

Not sure if the game lets you be reported to the SEC. Maybe you're a member of congress?

That should be a feature in Starfighter. 'tptacek?

If you play through the trading levels of Stockfighter, you'll become... intimately acquainted with the Committee for Exchange Security (CES), the in-universe SEC expy.

Can't wait to get through the experience!

Let me see how easy it is to implement in AVR assembly.

You could add a special opcode for that, something like SCCB (Short-sell and Call CES if Borrow).

Where's the short option?

In short term trading your profit/loss will be probably random. Saying that some day trades make money is like saying that some people might flip a coin and hit heads 20 times in a row. For a big enough group of people it will happen. Unless you are high frequency trader - then you make money every single day.

> Unless you are high frequency trader - then you make money every single day.

Only if you're the fastest HFT that day, Dodd Frank and the sub penny rule devastated HFT trading by now allowing them to bid better, so whoever is first in line wins.

That was fun - almost $32,000,000 from the first run. Made a killing on DELL (260.9% vs -61.1%).

If you keep playing, it's the same stocks again and again - would be even more fun if there was some pseudo real time/more recent charts too.

Along these lines, can anyone recommend a good tool to test real strategies against the real stock market, taxes, fees, etc included? Not looking for free necessarily, just something that I could actually learn and test against real world data with the same variety of options, futures, paper, forex, etc.

Also, I remember working on a Bloomberg machine at one point, and the license was something like 200k a year, and I was told most of that wasn't for the machine but for the level of data access. At what point does the data become free?

Have a look at quantopian.

Wow, thanks for the suggestion, that is extremely interesting how they are doing things. Crowd source hedge fund where you can work on your own algo's...

Now the only problem is my curiosity is piqued and I feel like I'm about to waste way too much time learning about this stuff.

this is fun, but it gives you clues on where it will go next as the frame shifts before big movements.

Meh, but the scale change is too quick to see if it's going higher or lower (at least to me) so I don't think it gives that much of an advantage.

Even trying to use that information I can't seem to get much better than the 50th percentile, on average. Which is pretty much what I had before I read your comment.

After about 30% of one round you can see from the previous trades if you are more on the low or more on the high side though. And the moment you see (or rather "feel") the scale change you can make a decision.

I couldn't pinpoint it as well as the comment above yours but I also feel the interface helped a lot in making good decisions.

Played through all the stocks, got up to 96,000.00$. Pretty fun, especially some of the stocks which tank (too bad there's no short option).

Works surprisingly well on mobile too.

Well definitely gets easier. Got up to 10+ million starting from 500, and then used that to get to 16 billion... One more round and I'm richer than Bill Gates lol...

And made it past a trillion...

This is a cool game! But does it represent the real world? It seems to be a game of reflexes. Or does ranking in the top 100 means I actually can predict the future?

> does it represent the real world

When you are done with the game the stock ticker appears in the top left of the chart - that was the real world price you were trading.

I am not sure they are taking into account realistic transaction fees.

It would be interesting to add standard fee's to the game.

They might be specifically picking stocks that have done well.

In that case, you would on average make more money in this game than you would in real life.

I think they chose well-known corporations; Enron was one.

I had Coke and Netflix.

In that case, the stock market seems like easy money, especially if it's about reflexes as you can have a computer do the work and trade in hundred of stocks simultaneously.

Is this game designed by Bloomberg to lure amateur day traders into their system?

The games is designed in such a way that it is so easy to make big virtual $$$$$$$.

Next step - actual trading account and $$$$$ losses (while Bloomberg collects commission fees).

I got #2 on AAPL on the first try. I ignored the scale and just bought based on dates: Bought asap, sold in late 1999, bought in late 2001, sold mid 2007, bought early 2009, sold May 2015, bought mid-August 2015.

Stopped playing after 170B. Too bad they don't mix up the charts, kinda fun. I couldn't quite figure out the scale shifting helping, seem more like a momentum thing.

So nicely done, would love to see the code on github thought

It's just Javascript. You should just download it to run locally.

In the real world you wouldn't always buy / sell at the price you hit, someone could front run or by the time your sell hits the market there could be lack of buyers and the price will have tanked.

So this makes it seem easy, when the trade differ in real life from what might look available one minute.

1. Click at the start

2. HODL til' the end of time

3. Profit

This goes the show how much of a gamble it is when investing by charts.

I made $1.6148008835081285e+21 LOL

The right axis scaling leaks a lot of future info...

Just buy at the start and hold until the end, you will outperform the market and most of the people.

It's always a different stock. If you do this with Enron, you will lose 100%.

I sold it in December 2000 in the game - 1650% return.

Of course I'm taking on an average behaviour

Ya buy and hold is the way to go in real life.

Just like in real life.

I got lucky - first try #9 on GOOGL, outperforming 99,4% of all players http://imgur.com/yWohi4J

and #8 outperforming 99.6% a few tries later http://imgur.com/75raZvr - returns 6462% on a stock that lost 75.3% over that period

Nice. I made $83,664 (9658%) trading $MCZ, beating the stock! And beating the market! I beat 100% of people trading $MCZ, ranking #3. http://i.imgur.com/dTaCufc.png

Reply on news.ycombinator.com