HelpMeManageRisk: 101

Wassup deboggers. As I mentioned, my goal is to create one of the best communities in the space. While you are here !debogging, I hope to bring you some new experience that can eventually make you a better/more sophisticated NFT/crypto player.

In this article, I will talk about why I take derisk metric so seriously in HMDT bot. Let's roll.

The NFT market is different from traditional financial markets in a few fundamental ways, but maybe we can learn a couple of things from those guys on the wall street. One such area is risk management, and that is what we will be covering today in this short risk management 101!

Most traders in the financial markets are risk seekers who are willing to be exposed to a certain level of price uncertainty in order to be compensated. At the very minimum a trader should wish to make a profit that is more than a risk-free investment, like putting your money into a savings account. That additional profit is called excess return, and an asset with potential to generate excess return can be measured by its risk premium.

The amount of health insurance you pay every month is a type of risk premium: if you are unfortunately injured or sick and got a large medical bill, your insurance can often cover part of it. The amount covered is essentially a return on your initial investment which the insurance company will pay to you. Maybe 99% of the time you are perfectly healthy and won’t get paid much by your insurance. But the other 1% of time when you get that huge bill, insurance might pay an amount that is several times of your annual premium. That 1% probability determines how much your monthly premium should be, among other things.

Since risk premiums can be priced by the market, they can also be bought and sold. Here are some common buy/long risk premiums trades:

· Long $100 strike call options of Gamestop when the stock is trading $5 a share

· Buy 10 Powerball lottery tickets

· Mint HMDT

And some examples of shorting/selling risk premiums:

· Lending $10,000 to someone you don’t know well in exchange for a 10% annual interest

· Selling life insurance

· Short $400 Tesla call options in Mar 2022 when the stock is trading $350 a share

You can make or lose money by either buying or selling risk premiums: after all, if you are right 100% of the time about if an asset is going to move up or down in price, you will make a shitload of money anyways. But for real deboggers like us who don’t have crystal balls and make mistakes all the time, it is important to know how these two different types of trades can help or hurt you in the long term.

Without getting too technical into the financial theories behind all this, let’s simplify things a bit and just consider these two games:

Game A: You pay $1 each round to play a roulette with a 1% chance to win, and the payout is $200. You get nothing the other 99% of time.

Game B: You own a roulette wheel and charge $3 each round for others to play. There is also a 1% chance your counterparty wins in which case you will pay them $200.

Game A is a long risk premium trade; Game B is short risk premium. Both games have an expected payout of $1 per round. If you have $100 initially and can play either game for 100 rounds consecutively, which game will you play?

----------------------------------------------------------------------

Let’s run a simulation to compare these two games. The chart below compares the cumulative expected payout for the two games, by averaging results from 10,000 trials. Unsurprisingly, both games have a very similar path of payout initially since the EV for each round is both $1. But as we play more rounds, Game B seems to break from the $1 per round linear growth, while Game A more or less still follows a straight line. That is because we have a limited amount of fund to start with: when you run out of the initial $100, you are simply busted and cannot continue. Game B has a much higher ‘death rate’ since you can lose $200 in any round with a 1% chance, but you can’t die in Game A within 99 rounds. That is just simple math.

One might argue that we designed the game to be biased towards Game A though, since there is zero chance to go broke in Game A within 99 rounds. What if we continue playing? This is what I am showing below in this second chart.

If you get unlucky in Game A and didn’t win a single time in the first 100 rounds, you will also be broke. And this is why you see an inflection point in the red line at round 100. And by round 1000 your expected payout for Game A is only around $600 instead of $1,000. But we are still seeing Game A outperforming Game B: In fact, the cumulative difference between the payouts only grow larger instead of flattening out or shrinking. The cumulative death rate by round shows the same story: at any of these rounds, you have a higher likelihood to be busted in Game B vs in Game A.

Round	Death rate Game A	Death rate Game B
100	0%	40%
200	36%	45%
300	36%	48%
400	41%	50%
500	41%	51%
750	43%	53%
1000	44%	53%

----------------------------------------------------------------------

This leads to the first suggestion I would give in this risk management 101: In the world of investment, the first and maybe the most important strategy is just to survive. In practical sense, most of us are bounded by a limited amount of money to invest and you don’t want to be in a situation where you might be broke over a few bad investments. Short risk premium trades can provide you a stream of good income in the good times, but can also get you ‘instantly killed’ in a bad scenario; Long risk premium trades, on the other hand, works in the opposite fashion. You might need to pay that premium most of the time, but that won’t easily get you close to being broke and there is a chance you can get a handsome reward back.

For asset management companies that have a much larger pool of fund to start with, seems like they don’t need to worry much about the limit on capital as much, right? But risk is still a huge concern if you are comparing these two types of games. Instead of a fixed amount of money, portfolio managers at these companies are usually limited to a certain VaR, or Value at Risk. The idea is that for each portfolio manager the company/fund can tolerate a loss of X dollars on a really bad day, which is usually measured by loss in a 99-percentile worst case, also called P99 loss (95, 90 also widely used). So the PM allocating all her money to treasury bonds can be allocated a much larger amount of money nominally, than her colleague who buys TSLA, Bitcoin and other more risky assets. But their VaR could be the same which is what really matters.

Back to our games. If you are a PM who finds 1,000 different Game A type trade strategies and another 1,000 Game B type strategies, how should you allocate your money to maximize return relative to risk? Suppose that the PM will make one bet on each strategy daily, which is equivalent to $1,000 expected return per day. Also assume that the strategies are uncorrelated with each other. The simulated P99 loss for Game A is -$200, while for Game B this is -$640. For Game A, the ratio between expected return and VaR/P99 loss is 5, which can be typically considered as a great opportunity; That ratio is only about 1.6 for B. In other words, normalized for worst case loss, Game A is 3 times better than Game B.

The no correlation assumption we made is critical too. If there is a significant positive correlation between the games, the P99 loss for Game B type trades can be exponentially worse. Say if you lend money to 100 different people in return for some interest payment, and you might think that these are 100 uncorrelated short risk premium investments, since these borrowers don’t have much in common. You think that each of them only has 1% chance of not being able to give the money back, so the chance of say 20 of them defaulting at the same time should be very low. Therefore, you can conclude that this is a pretty low risk investment right? Well, if you are familiar with recent financial history, this is exactly what leads to the 07-08 subprime mortgage crisis…

And here comes the second suggestion for this risk management 101: No matter how good you think the expected return on a trade might be, always normalize by the worst case loss. If you think buying 100 BTC can make you $1M dollars in the next year, that’s great; But what would be the worst case? Could it be losing $1M, or even more? This is the kind of question you will need to ask yourself, and you need to be truthful to yourself. Don’t understate the risk, especially when you don’t know what is a good answer. Too many bad trades are coming from people who confused ‘I don’t know what is the risk here’ with ‘I don’t think this is risky’.