Recently, there are some concerns in auction duration and bid arrivals in study of online auctions. In online auctions, seller selects duration, and the bid arrivals pattern is deemed to the indicator of bidding strategies and behaviors of bidders. The best known phenomenon in bid arrival is the late bidding, which describe the following bidding strategy used frequently by many bidders: bidders tends to place bid as late as possible. Noticed there is a void in growing body of literature on online auctions in comparison study of participants' behavior across similar but different websites, we sampled eBay auctions from several global websites simultaneously and compared the duration distributions and bid arrivals patterns across these websites. The statistical results indicate that the comparison study is meaningful. Finally, we imply that more advanced data mining tools are needed in analysis of online auction data in further research.