Two Hundred Versus Eight

Two documents both match the query term kubernetes. The average document length in the index is 500 terms.

doc	term frequency	document length
S (keyword-stuffed)	200	2000
T (real tutorial)	8	400

BM25 scores the term frequency part as

$\frac{tf}{tf + k_1\left(1 - b + b \cdot \frac{dl}{avdl}\right)}$

with $k_1 = 1.2$ and $b = 0.75$. Both documents get the same IDF. What is the ratio of S's term frequency component to T's?