python - Two factor model with kalman filter for Brent oil forward curve -

i implemented 2 -factor kalman filter python recently. got output hard tell if doing right or wrong. model complicated , impossible me put codes here. hoping if there did before can give me idea if reasonable following result...blue line actual forward curve , red line kalman filter

i satisfied shape. however, while doing on rolling basis observing prediction higher actual forward curve(1-2 dollars on average). common problem using kalman filter? can improve accuracy?

this question might seem ambiguous believe tried kalman filter before more or less have idea getting here. appreciate if can provide suggestions should regarding output.