There is considerable variation in recommendations for optimum tryptophan (Trp):lysine (Lys) ratio in diets of starter pigs up to 35kg body weight (BW) which may partly be due to differences in the composition of experimental diets, the experimental approach, and the mathematical model used for data analysis. The objective of this study was therefore to determine the Trp:Lys ratio required for maximum response of pigs (15–35kg BW; BW) offered a wheat–barley–soybean meal (WB) or a corn–soybean meal (C) based diet. Three mathematical models were applied to test to which extent the choice of model will affect the estimate of optimum Trp:Lys ratio.A 4-week dose–response growth study was conducted with 880 crossbred castrated male and female pigs. The pigs were randomly allotted to eight dietary treatments per diet type. A Trp-deficient diet (T1) was formulated for each diet type that met animals׳ requirements for all amino acids other than Trp (1.4g standardized ileal digestible (SID) Trp/kg diet) and Lys (10.5g SID Lys/kg diet). L-Trp was added to T1 diets to create six additional diets per diet type (T2–T7) with SID Trp concentrations increasing from 13.1% to 19.7% in WB based and from 14.3% to 20.9% in C based diets. To test whether Lys was the second-limiting amino acid, a Lys adequate diet was formulated (T8) for each diet type by adding crystalline Lys to the respective T7 diet. Pigs had free access to feed and water. Individual BW was recorded every two weeks and pen feed intake daily.For both diet types BW gain increased (P<0.01) and feed to gain ratio improved (P<0.05) with increasing Trp:Lys ratio. In the Trp-deficient diets feed intake was reduced by 100 and 60g/day in WB and C based diets, respectively, compared to diets 2–8. According to the broken-line, curvilinear plateau, and exponential models, highest BW gain was reached at Trp:Lys ratios of 16.8%, >19.7%, and >19.7% for BW based, and 18.0%, >20.9%, and >20.9% for C based diets, respectively. The respective ratios for the feed to gain ratio were 17.0%, 18.5%, and >19.7% and 19.9%, >20.9%, and >20.9% for WB and C based diets, respectively. The results confirm published ratios determined in WB based diets and indicate a slightly higher optimum ratio in C based diets. However, the mathematical model used for data analysis, the chosen response parameter, and the target performance level strongly affect estimates of the optimum dietary Trp:Lys ratio.