You can use several different kinds of data to make predictions about the future. Some of these are based on what people think, some on what people do, and some on what they say.

The most important kind is based on the way people behave when they are faced with uncertainty. When there are many competing things you could do, and no one can tell which will be best, the best thing to do is nothing. And that is what we see in markets.

A good cryto currency has a lot of competition; it needs to be very volatile because buyers and sellers must always be moving, or they won’t trade. It also needs to have very low transaction costs; if you expect lots of transactions, then you need to spend a lot less money or time than ordinary currencies, so you need to pay for your transactions with something that does not have an associated transaction cost.

The time and effort it takes to choose the right cryto currency is considerable. Here’s a brief guide that can help.

You need to consider:

How much work will have been invested in the cryptocurrency before its launch? How easy is it to calculate the value of this investment? Is the cryptocurrency backed by something with intrinsic value, like gold, or by a company with a predictable business model, like Amazon or Google? Will people be willing to use the currency in the future? Is it likely to become more valuable over time?

Predictive metrics are quantitative tools that measure a thing’s past performance, and use past performance to predict future performance. For example, price is a predictive metric. If you know the price of something in the past, you can make an educated guess about how its price will be in the future.

The crypto currencies are all based on cryptography, which is like code but with a lot more math. The code is secret; you don’t know what it does, but if you break it, you can’t read what it does. But if everyone knows everyone else’s codes, it’s hard for anyone to break them.

Early crypto currencies used regular encryption algorithms: the same kind of math that banks use to keep credit card numbers safe from hackers. They were called “proof-of-work” or “scrypt” currencies because they had to solve a computation problem before they could be spent. The problem was easy enough that it could be done by a computer running an ordinary program (not by hashing), so there was no reason to think anyone would bother trying to crack it; that meant there was no need for a lot of people working on cracking it, and no need for the cruder kind of mining that miners now do.

But as soon as mining

Crypto currency is a new technology, but there are already some good general principles you can use to pick a winner.

First, pick a currency that has the biggest user base: the dominant coins in the market. The bigger the user base, the more likely it is that people will be using it as money. Second, pick a currency that has low volatility: a price change of 5% or less matters less than one of 10%. Third, pick a currency that has lots of use cases: not only does this make it more useful in day-to-day transactions, but in itself it gives it more value.

One way to think of crytos is as a form of insurance. If you own one you can pay for things in it, like a mortgage; but if you don’t own one, someone else does. Crytos are like insurances against bad times: you can pay now, or pay later. But unlike most insurances, crytos are also good investments: if the price of your cryto goes up, you can sell it for more than you paid for it. That’s why these investment-grade securities are called “securities.”

Crytos aren’t perfect insurances against bad times. Yes, by holding one you reduce the risk that something bad will happen to your money. But they don’t reduce the risk that something good will happen to your money. You still have to pay taxes on it and look after it yourself. Cryto ownership also reduces the risk that governments won’t do what they’re supposed to; but that is part of any security, regardless of its kind. Crytos just make this kind of security cheaper and easier to buy and sell.

The most important thing to think about is that there is no such thing as a free lunch. There is no such thing as a free currency, either.

That doesn’t mean you should never invest in a currency. It does mean, though, that you have to think about the costs of doing so. Crypto currencies can look pretty great; but if you don’t have anything else to put your money into, they are not free lunches.

Crypto currencies are still new; I am writing this shortly after Bitcoin Cash forked off from Bitcoin in August 2017. At the time of this writing, I am unsure if it will be better than Bitcoin or worse; even so, it’s too soon to tell which one will be better.

In any case, the question isn’t whether or not to invest in crypto currencies; it’s only how much to invest.

In the early 1990s, the International Monetary Fund (IMF) set up a network of 20 computers around the world. It called it “the World Bank for the 21st century.” The idea was that if the computers made lots of different bets on currency movements, they would come up with a good guess about what would happen to currencies in real life, and how much worse or better they would do when they were actually trading.

It was an important attempt at a new kind of forecasting. For example, no one had ever tried to predict stock prices before. So how did they do? The answer is: pretty well. In fact, the predictions were so accurate that they came close to predicting changes in stock prices that were already occurring. They beat four leading big-data companies by 25%.

The IMF eventually shut down its Web site and sold off its computers. But in our own work we have developed something similar. If you give us a prediction of current events and we make some bets based on it, will we beat regular forecasts? We think yes.