Why NewsFutures Sucks – A disgruntled trader’s story

Via reader Jill Peters, a long and detailled blog post from an unhappy NewsFutures real-money trader from Holland. Let’s see whether his criticism is fair —if it’s possible for a third party to judge (we will see). Below is a collage of excerpts; emphasis are added.

—

Why NewsFutures sucks

In the past few months I’ve played on the Politieke Aandelenmarkt (PAM; Political Stock Market), which was organised by Dutch newspaper De Volkskrant to predict the results of the Dutch general elections yesterday. The system was created by Franco-American company NewsFutures. Even though the game has not yet ended, one conclusion is inescapable: NewsFutures sucks. Its game system designers lack imagination, and its programmers are incompetent idiots. If you want a predictive stock market, go to one of its competitors.

On the PAM we play with real money. I invested € 100 (the maximum), as did most of the other traders. BTW: at the time of writing I’m up to about € 205, or a 105% ROI, so I don’t write this post because I lost money.

—

#1. Game System

This sounds great, but there’s one flaw: in yesterday’s elections two parties overshot their caps by a significant margin. In itself this is not NewsFutures’ fault; De Volkskrant should have foreseen this eventuality and adjusted the caps. Nonetheless, the whole situation was caused by the lack of imagination of NewsFutures’ game designers, who’ve only looked at the US situation.

It’s even worse at the coalition market. There are no caps, and until yesterday there were quite a few low-probability coalitions that I’d have liked to sell short on, but couldn’t because of the ridiculously high deposit.

In short, NewsFutures’ system is created with a simple market in mind (say the US or UK elections) where there are only two or three parties that really matter. In such a market it works fine; in more complex markets it doesn’t. []

#2. Back-End and Front-End

While playing on the stock market, I noticed several flaws in NewsFutures’ back-end and front-end systems. Occasionally an order would go astray; once I bought a few shares, and though the money was immediately deducted from my account, the shares were not added to it. This was solved a few minutes later, but it would be nice if such errors didn’t occur at all.

When it comes to front-end (my [specialty], after all), NewsFutures is reasonably sucky. They’ve never heard of web standards, usability, or accessibility. Two examples will clarify this. []

#3. Problems on Election Night

To my surprise, the system kept up quite well. Although it clearly couldn’t keep pace with incoming orders, neither did it go down entirely. In itself, NewsFutures’ trading system withstood this challenge.

Something else went wrong though; disastrously so.
—
Without giving my user name and password, I was logged in as another player! Although I immediately went back and entered my user name and password, on the forum I saw that plenty of other traders had experienced the same problem. []

—

My Take: The testimony of this Dutch real-money trader is striking.

- The NewsFutures team that designed the Dutch politics prediction markets did not fully anticipate all scenarios.

- The NewsFutures software, clearly, was not up to the task for real-money prediction markets.

- I have been covering the field of prediction markets actively since 2004 and I will tell you something. BetFair is the only reputable company in this business. It’s a company built around its technology and traders. My readers will have noticed that each time that I write that, I attract critics who castigate BetFair as “gambling”. [*] In my view, we have a big problem in the prediction market industry —many people refuse to understand that information technology is the key, that it is costly, and that, in this low-cost industry, only a trio of market leaders have enough money to pay for a high-quality IT infrastructure.

Some have suggested that exchanges organizing socially relevant prediction markets should be seen as think tanks. This is the Earth’s dumbest idea. Think tanks don’t deal with technology and customers. The promoters of such lofty ideas should take an appointment with Reality —capital “R”.

—

Addendum: [*] On top of receiving e-mail insults from Ireland.

—

Addendum 2:

Emile Servan-Schreiber of NewsFutures has posted a comment on the Dutch trader’s story:

Yes, election night in Holland, Nov 22nd, revealed some technical limitations in the political market we set up for de Volkskrant. There’s nothing like a real-life massive stress of a piece of software to discover where it falls short of perfect. Then, you pick yourself up, solve the problem and move on. Had Friedrich Nietzsche been a software developer, he would have agreed that the bug that doesn’t kill you makes you stronger. As a result, our software is better now than it was three weeks ago. Here are some facts to keep this “debacle” in perspective:

1) Despite the technical bumps on election night, no one has complained that they have unfairly lost any money as a result. Under massive pressure, the software did not go berserk but demonstrated rather graceful degradation. (And with the consequent fixes in place, it will behave even better the next time it faces a similar situation.)

2) After venting his frustration, the “unhappy trader” has kept on trading rather happily: 472 transactions in the three weeks since election night. If you did not know better, you’d think this guy still has some confidence in our system.

Cheers,

–Emile

—

Addendum (December 14): Deep Throat on the NewsFutures’s Dutch debacle