The cost of reworking and finishing steel casting parts show a great influence on the economic efficiency. Hot tears, which appear as cracks close to the surface are the most common casting defect in steel casting. The repair work of these hot tears costs a lot of money and time, whereas the exact mechanism of how these cracks appear remains still unknown. Next to the influence of the geometry of the casting part, metallurgical influences such as the casting temperature above the liquidus temperature and the chemical composition of the alloy are under suspicion to increase the hot tearing tendency. While the influence of the geometry was investigated multiple times in the past, a systematic examination of the influence of the metallurgy on the hot tearing volume was just performed for the continuous casting process and for steel casting so far. The goal of the present work is the study the influence of the metallurgy on the hot tearing volume, by a comprehensive process data analysis and by performing a carbon variation with dipping tests of 3 different alloys. Two of these alloys were casting steel grades of the G17-group and the third melt was a Fe-C-1%Mn alloy. To predict the connection of the hot tearing volume and the process data, so called “quality-prediction-indexes” were calculated based on the chemical composition of each cast by a micro segregation model. One of those indexes was the thermal strain in between the mushy zone. Another calculated quality index was the disturbed strengthening, which refers to the start of the austenite formation in the mushy zone. The calculation of the “quality-prediction-indexes” is dependent on the thermodynamic databases (ThermoCalc-TCFE8.1 und FactSage-FSstel 2015). Therefore these databases were validated by high temperature DSC-measurements. Low alloyed casting steel grades such as G17CrMo9-10, G17CrMoV 5-10-2 and G17CrMo 5-5 are described exactly by the used databases. The phase transitions of alloys with a high chromium content such as Gx4CrNi 13-4 and Gx5CrNiCu13-3-1 and very complex alloys such as GX12CrMoCoVNbNB9-2-1 and Gx8CrWCoVNbNB 9-3-3 can’t be calculated accurately at the moment with the present databases. Both the quantity and quality of data was quite limited for a statistical analysis so the results show no a clear tendency. On one hand there was a strong influence of the model and the alloy group as well as a positive tendency for an increased casting temperature. On the other hand, the calculated “quality-prediction-indexes” showed only some statistically hedged tendencies. This might be caused by the strong influence of the model. The dipping tests of the Fe-C-1%Mn model alloys confirmed the results of the literature and a new systematic and digitized evaluation algorithm was developed in the present work. The examination of the carbon variation of the G17CrMoV 5-10-2 casting alloy showed a shift of the maximum of uneveness closer to cB in the hypo-peritectic range.