x <- runif(10)
y <- x - x**2 + rnorm(10, sd=1)
M <- lm(y ~ poly(x, degree=9))

sum((y - predict(M))**2)

[1] 2.889376e-31