They are comparing the daily trading net revenues to VaR to see if and how many times 95% VaR was exceeded. The results were “Trading losses incurred on a single day exceeded our 95% one-day VaR on two occassions during 2010” and in no occassions in 2009.
Now, they don’t give one-day VaR for every day in 2010, but in another table in the same BB they say the average daily VaR in 2010 is 134 Mil. In the last graph in BB there are 3 days with more than 100 Mil in losses. One can assume that the two occassions during 2010 where one-day VaR was exceeded would be in 2 of these 3 days, but without having the daily VaR and daily trading net revenues data, you cannot be sure which two occasions they are talking about in BB.