Voevodsky defined the slice filtration on the motivic stable homotopy category $SH(S)$ over a Noetherian scheme $S$. In the article [Open Problems in the Motivic Stable Homotopy Theory, I][1], Section 2, he defined $SH^{eff}(S)$ to be the smallest triangulated subcategory in $SH(S)$ which is closed under direct sums and contains suspension spectra of spaces. It follows from a result of Neeman that the inclusion functor 
$i_n:\Sigma^n_TSH^{eff}(S)\to SH^{eff}(S)$ has a right adjoint $r_n$. Since $i_n$ is full, we have a natural isomorphism $\mathrm{id}\simeq r_n\circ i_n$. Denote $f_n=i_n\circ r_n$. Applying the counit $f_{n+1}\to\mathrm{id}$ to $f_n$, we get a natural transformation $f_{n+1}\circ f_n\to f_n$. Voevodsky claims that $f_{n+1}=f_{n+1}\circ f_n$ so that we get a natural transformation $f_{n+1}\to f_n$. The slice functor $s_n$ is defined to be the cofiber of this map. Denoting the inclusion functor $\Sigma^n_TSH^{eff}(S)\to \Sigma^{n-1}_TSH^{eff}(S)$ by $j_n$, (so that $i_{n+1}=i_n\circ j_{n+1}$), I calculated that 
$$f_{n+1}=i_{n+1}\circ r_{n+1}=i_n\circ j_{n+1}\circ r_{n+1}=i_n\circ\mathrm{id}\circ j_{n+1}\circ r_{n+1}$$
$$\simeq i_n\circ r_n\circ i_n\circ j_{n+1}\circ r_{n+1}=f_n\circ f_{n+1}$$
but don't see why  $f_{n+1}=f_{n+1}\circ f_n$. Note that $f_{n+1}=f_n\circ f_{n+1}$ also induces a natural transformation $f_{n+1}\to f_n$.

Am I missing something here? Or is there a typo in that paper?


  [1]: http://www.math.illinois.edu/K-theory/0392/newopen.pdf