About loot boxes

Just before the release of Star Wars Battlefront 2, a lot of debates took place about the loot boxes system in multiplayer mode.

I listened to “Silence on joue” this week, a french podcast about video gaming, and from what I understood, players were complaining that they would have to buy loot boxes in order to have access to some items, and those loot boxes would drop items randomly, so you could spend a lot of money before getting the item you’re looking for. They were also complaining about the pain-to-win model, but it’s not what Im interested in here.

What stroke me is this question: how can you tell it’s random? Items have different drop-rates, you click to open your boxe, cross your fingers and if you’re lucky you get what you want. From a gamer point of view, from an individual stand-point, it looks like it’s random. It feels like random. But is it?

Im challenging this based on reading Game makers are profiling players to keep them hooked and Behind the addictive psychology and seductive art of loot boxes.

As machine-learning gets bigger and bigger in profiling gamers, I don’t see any reason why the output of these algorithms could’t be also used to influence loots one get from loot boxes.

So, is it random? Is it AI? Will you be spending money to get something you’re doomed to never loot because some AI for some reason think it’s better not to let you have this?

From an individual stand point, you can’t make no conclusion. It’s impossible to tell. But assuming it’s random because it’s always been is worst for me than being left with an open question.

AI is sold to us as something that people will benefit from. It’s all marketed this way. But with this loot boxes thing, would AI be there to enhance my experience as a gamer or would it have been programed with another goal in mind, like maximizing revenue?

 

 

Flash Boys

It’s a great book by Michael Lewis about High Frequency Trading. It took me ages to start reading it, and I regret it because once open, you barely can’t stop reading.

It always seems easier to me to start with an example, so here’s what was happening before HFT: let’s say you’re a trader and you want to buy stocks. On your computer, you can see stocks available and their prices. Let’s say you want to buy 100 shares of XYZ that are available on the market for $10 each. You press enter, and it’s done.

One day, trader Brad Katsuyama discovers that when he tries to buy 100 shares for $10, it no longer works. The system cannot fulfill his request. He gets 90 shares for $10, but the 10 missing shares are no longer available at $10 on the market and he needs to pay extra money if he wants to buy them.

This wasn’t only happening to Brad Katsuyama. All traders were experiencing the same weird behavior from the market. It even got worst over time, with an increasing percentage of stocks they couldn’t buy at the original price.

To understand what was happening, there are 3 factors to take into account:

  1. The market doesn’t exist as an unified place. The market is made of exchanges. When you buy 100 shares, you don’t necessarily buy them from one exchange only. Your order can be fulfilled by several exchanges: one will have 15 shares, the second will have 30 and a last one will have 55, so you get 100 at the end.
  2. Exchanges are no longer these floors with lots of people and lot of noise, like in the 80s. Exchanges are digital now. Exchanges have servers, where transactions happen. And these servers are hardware devices. They physically exist, and as a consequence are located somewhere.
  3. There’s a law in the US which enforces brokers to buy the cheapest stocks first. As a broker you can’t buy 100 shares at $10 if at the same time there are cheaper options. You’ll start buying at $9.99, and once no more share is available at this price, you can buy for $10.

Based on this, some companies decided to get rid of the old telecom network and to build their own optical fiber network from New-York to Chicago. Their goal was to get a latency way better than what brokers were having. Say hello to HFT firms.

Now let’s get back to our example to understand what happens.

Before HFT, when Brad clicks enter on his keyboard to buy 100 stocks, the order is spread through wires. The signal arrives first to the closest exchange, which means the closest server, where he can buy 15 shares. The signal goes on, until it reaches the server of the second exchanges and buys 30 shares, and continues until the 3rd server where it buys the last 55 shares. As humans, it seems to us that all of this happens at the same time, because all of this take less than a second to be performed. But we’re wrong. There is an order in which things take place.

Now, with HFT, when the signal reaches the first exchange, it turns out that HFT is the first entity it will have to buy from, because HFT companies have small amounts of stocks priced just below the market they use as baits. They take advantage of the regulation to know that you are interested in buying 100 shares, so at the same time that they’re selling to Brad 1 or 2 shares, they send their own signal to buy 85 shares to the other exchanges. Thanks to the speed of their network, their signal reaches the second and third exchanges before Brad. When Brad arrives, the shares are no longer available at the original price, and he ahas to buy them for more.

HTF companies never took no risk on the market. They could simply not lose. One guy’s testimonial is crazy: over 3 years, he said there wasn’t a single day where they’d lose money. And though the advantage they had could seem unfair, it wasn’t illegal at all.

I hope you understand that this is really a basic abstract. The book goes deeper than this. But at least you get the point with HFT.

What I think Ill remember from this is how, when it gets to digital, we tend to forget the physical world.  I often smiled seeing stickers saying “There’s no cloud, it’s someone else’s computer”. We’ve got kind of the same idea here, with a more dramatic background I guess. Also, and though they are depicted as the baad guys, I must admit Im impressed by how smart the HFT guys had to be to come up with their idea in the first place.

 

Differential privacy

I know Im late but I heard about differential privacy this week for the first time. The idea is to introduce noise in the data so as to make it impossible to find out afterwards who answered what to a question, without damaging global results.

Let’s say you make a study about drugs and want to question people about their habits. Instead of just asking them if they do, you ask them to flip a coin. If they get head, they honestly answer the question. If it’s tails, they flip it again. This time, if it’s head they answer yes, and if it’s tails they answer no.

With this method, it’s impossible to know precisely who does drug, because 25% of people answered yes just because of the coin, so any individual can pretend he said yes because of the coin. But on a larger scale, you can still remove non relevant answers and deduce the average number of users in the population.

If 100 people take this test and you get 35 yes, you can assume that 20% of people do drug ((35-25)/50).

It seems that Apple is using this technique a lot, and so does Google and I guess a few others. If I understand the concept of how it works in social studies, it’s stil unclear to me how the tech and data industries leverage it. Not what we really win down the road.

 

Here

is the beginning.

I wanted a place to keep track of my thoughts and to tell stories about the world from my window. Or from my screen. Probably mainly from my screen.

I also tend to forget those stories and thoughts. That prevents me from having new ideas by connecting unrelated old ones together.

Let’s make it clear, Im mainly gonna talk about digital and tech. This should be my focus here.

My name is Nicolas.

38.

Based in the US.

I have a marketing and business background in robotics and software companies.

Welcome.