If I'm reading it correctly, Table X (page 2633) of Schwert (2000) (Journal of Finance, Hostility in Takeovers: In the Eyes of the Beholder?) says that about 78 percent of deals in 1975-1996 were successful.
However, this measure is constructed based on the acquisition of the firm, not the bid of the acquirer, so that if there are multiple bidders this is coded as a 1 if any succeed and a 0 only if all of them fail. So it isn't what percentage of bids succeed but rather what percentage of targets are ultimately acquired. I couldn't find any measure of the probability of a bid succeeding but it must be significantly lower.
One tricky issue is that while successful bids must be revealed unsuccessful one need not be, so there is selection bias in the sample of bids. Therefore, a narrower way of characterizing this result is to say that of the potential deals where at least one bid was made public 78 percent were ultimately acquired by some firm.