With the increased use of low ignition quality fuels in advanced compression ignition engines, the extended ignition delay and two-stage ignition behavior shown on the measured in-cylinder pressure profile raise a question about at what point of the pressure trace should be identified as the start of combustion (SOC). Previous studies used numerous methods, but a systematic evaluation is lacking, particularly for low ignition quality fuels used in a small-bore engine. The present study bridges this gap by performing high-speed imaging of OH* chemiluminescence in a small-bore optical compression ignition engine, against which various methods of ignition delay calculation are assessed for a correct representation of the start of high-temperature reaction—i.e., the actual SOC. These methods include the SOC defined as the pressure recovery point, zero-crossing point of the peak pressure-rise slope, 50% peak pressure-rise point, peak points of the second-order pressure derivative and the change of apparent heat release rate (aHRR), and 10% heat release point (CA10). Three custom-made fuels with a cetane number of 30, 40, and 51 are used with all the fuel physical properties almost fixed but the cetane number. The results show that the first signals of OH* chemiluminescence are measured closest to the pressure recovery point while the other methods return a later SOC or longer ignition delay time. Given that the ambient gas temperature is not fixed but changes during the ignition delay in an engine, a correlation between the ignition delay time and calculated bulk gas temperature is also evaluated for a range of temperature points, including the temperature at the start of fuel injection (SOI), temperature at the SOC, mean temperature during the ignition delay, mean temperature between SOI and the end of injection (EOI), mean temperature between EOI and SOC, and temperature difference between SOI and SOC. A correct correlation is found only with the temperature at SOC and the temperature difference between SOI and SOC.