Obviously it's a question of timeframe. The %Profit of ~30% across brokers in that table is for a single year. If you have a 70% chance of loss any given year, then over a 5- 10 year timeframe your chance of loss tends toward 100%. Hence the '95%' number you often hear.
Actually, the report seems to describe a single quarter. So for IB, if we assume independence for each quarter and for each trader, then annualizing gives:
1 - ( .465 ^ 4 ) = .953
i.e., a 95.3% failure rate (and that's for the best brokerage in the list -- the worst in this list would score, by the above extrapolation, an average yearly failure rate of 99.7%!). Of course, each quarter is not independent for each trader, so the real failure rate is probably significantly lower -- but it still seams reasonable to assume that the failure rate accelerates when stretched out over time, and many other analyses (including published academic papers) seem to confirm this.
That said, I think it's also true that the oft-cited "95% failure rate" is more of a rough approximation than a hard statistic. Surely quite a few traders make money almost every year and dramatically outperform the market, and this too seems to be backed up by the few academic papers that have data at individual-trader granularity.
Also, any given broker probably has quite a few newbies who trade actively for a few weeks, then drop out, vs. some veteran traders who are consistently profitable and stay for years. The differences in time frames, capitalization, and experience muddies the meaning of "failure rates" for trading -- it's not nearly as clear as more familiar stats such as MLB batting averages.
I'm not surprised that the stats are particularly bad for retail forex where the broker sets the spread rather than allowing their clients to trade in a competitive market (e.g. forex futures on the CME or an ECN with competitive setting of the bid/ask). If the spread is not set competitively, I'm sure people are paying far more spread on every trade, and that really adds up over time. This is also (to a lesser degree) a problem in equities, where there is competitive setting of bids/offers on the exchanges, yet some players can jump the queue through "broker internalization", "broker priority", or "sale of order flow to wholesalers".