I wonder if anyone is using the benchmark cases in "Standard securities calculation methods" issued by Securities Industry Association (Vol 1, 3rd ed.) to calibrate their implementations for US Treasuries. My impression is that this is still an important reference in the industry. I have some problems to verify a few cases (ex. #18) and would really like to discuss this with someone. Thanks.
Edit: adding details on example 18
case #18A Benchmark result for price is 90.422798. My result is 90.422450. Small diff, yes, but these are benchmarks to verify an implementation so need to match all decimals (providing benchmark is ok which I actually doubt it is after spending some time with this). FYI, it is a case with odd first and odd last periods.
Type: Treasury bond Settledate: 10/12/92 Maturity: 06/15/20 Issuedate: 08/01/92 Fistcpndate: 12/31/92 Lastcpndate: 12/31/19 Daycount: ActAct Freq: semi-ann Cpnrate: 5.75% Yield: 6.5%