Both attempts are wrong.
This reaction is far too hot to be complete. It will not produce just CO2 and H2O like combustions in abundant air do, but lots of CO, plus H2, KOH, and many more.
The proportion of CO vs CO2 is uneasy to estimate by hand because other compounds add their mess. Software is the only reasonable way to go, for instance Propep or RPA.
Because much CO is obtained, the "best" proportions of reactants are not 5:48 moles neither.
This is what Propep computes from 170:100 mass oxidizer:fuel to find 1454K
but 290:100 would attain 1635K.