On a conjecture of Andrews and Bachraoui Koustav Banerjee, Kathrin Bringmann, William Keith Abstract . Recently, Andrews and Bachraoui considered a generating function F k , m ​ ( q ) F_{k,m}(q) associated with certain two-color partitions, and conjectured that this function has non-negative coefficients for m = 1 m=1 . They showed this property for 1 ≤ k ≤ 4 1\leq k\leq 4 . In this note, we prove that F k , 1 ​ ( q ) F_{k,1}(q) has non-negative coefficients for 5 ≤ k ≤ 10 5\leq k\leq 10 . Moreover, we show that, as k → ∞ k\to\infty , F k , 1 ​ ( q ) F_{k,1}(q) is related to Ramanujan’s third order mock theta function ω ⁡ ( q ) {\omega}(q) and to quotients of certain q q -binomial coefficients. 2020 Mathematics Subject Classification 05A17, 11P81. 1. Introduction and Statement of Results Throughout let (a)_{n}=(a;q)_{n}:=\prod_{j=0}^{n-1}(1-aq^{j}) for n\in\mathbb{N}_{0}\cup\{\infty\} . Moreover let (a_{1},...,a_{\ell})_{n}=(a_{1},\dots,a_{\ell};q)_{n}:=(a_{1})_{n}\cdots(a_{\ell})_{n} for \ell\in\mathbb{N} . Andrews and Bachraoui [ 3 , Conjecture 2.8 and Conjecture 2.9] conjectured that for k,m\in\mathbb{N} and for k=2,m=4 , the coefficients of the q -series \sum_{n\geq 0}\frac{\left(q^{2n+2},q^{2n+2k};q^{2}\right)_{\infty}}{\left(q^{2n+1};q^{2}\right)^{2}_{\infty}}q^{m(2n+1)}=:\sum_{n\geq 0}c_{k,m}(n)q^{n}=:F_{k,m}(q) are positive. Note that for m=1 , we have (1.1) F_{k,1}(q)=\sum_{n\geq 0}\frac{\left(q^{2n+2},q^{2n+2k};q^{2}\right)_{\infty}}{\left(q^{2n+1};q^{2}\right)^{2}_{\infty}}q^{2n+1}. Andrews and Bachraoui [ 3 , (2.2), (2.3), (2.5), Theorem 2.6] proved that c_{k,1}(n)>0 for n\in\mathbb{N}_{0} and 1\leq k\leq 4 . Moreover they proposed the following conjecture (see [ 3 , Conjecture 2.8] ), which we partially prove in this paper. Conjecture 1.1 . For k\in\mathbb{N} and n\in\mathbb{N}_{0} , c_{k,1}(n)>0 . To state our theorems, for A(q)=\sum_{n\geq 0}a(n)q^{n} we mean, by A(q)\succeq 0 (resp. A(q)\succ 0 ) that a(n)\geq 0 (resp. a(n)>0 ) for n\in\mathbb{N}_{0} . Theorem 1.2 . For k\in\{5,6,7\} , we have F_{k,1}(q)\succeq 0 . Using a computational approach we extend this to k\leq 10 . Theorem 1.3 . For k\in\{8,9,10\} , we have F_{k,1}(q)\succeq 0 . Our next result relates F_{k,1} to the mock theta function \omega of Ramanujan [ 5 ] (1.2) \omega(q):=\sum_{n\geq 0}\frac{q^{2n(n+1)}}{\left(q;q^{2}\right)^{2}_{n+1}}. Theorem 1.4 . We have \lim_{k\to\infty}F_{k,1}(q)=q\omega(q). In the following theorem, we obtain more information about the function that prevents F_{k,1}(q) from agreeing with q{\omega}(q) . To state our result, for A(q)=\sum_{n\geq 0}a(n)q^{n} and m\in\mathbb{N}_{0} , define \operatorname{coeff}_{\left[q^{m}\right]}A(q):=a(m). Theorem 1.5 . For k\in\mathbb{N} , we have q\omega(q)-F_{k,1}(q)=q^{2k+1}E_{k}(q)\ \text{with}\ E_{k}(q)\in\mathbb{Z}[[q]]\ \text{and}\ \operatorname{coeff}_{\left[q^{0}\right]}E_{k}(q)=1. Theorem 1.5 immediately implies the following. Corollary 1.6 . For k\in\mathbb{N} and 1\leq m\leq k , we have \operatorname{coeff}_{\left[q^{m}\right]}\left(q\omega(q)-F_{k,1}(q)\right)=0\ \ \text{for}\ \ 1\leq m\leq 2k. The remainder of the paper is organized as follows. In Section 2 we recall some standard q -series transformations and define notation. In Section 3 we transform F_{k,1}(q) into certain forms which assist in proving Theorems 1.2 , 1.4 , and 1.5 . In Section 4 , we prove Theorem 1.2 and Theorem 1.3 , showing the positivity of the coefficients of F_{k,1}(q) for 5\leq k\leq 10 . Next, we prove Theorem 1.4 , and Theorem 1.5 , exhibiting the relations between F_{k,1} and the third order mock theta function \omega . Finally, in Section 5 , we presents a few plausible approaches to address 1.1 . 2. Preliminaries Recall Heine’s transformation [ 4 , Corollary 1.2.4] . Lemma 2.1 . For |q|,|t|<1 , 0<|c|<|b|<1 , we have \sum_{n\geq 0}\frac{(a,b)_{n}t^{n}}{(q,c)_{n}}=\frac{\left(\frac{c}{b},bt\right)_{\infty}}{(c,t)_{\infty}}\sum_{n\geq 0}\frac{\left(\frac{abt}{c},b\right)_{n}\left(\frac{c}{b}\right)^{n}}{(q,bt)_{n}}. By [ 1 , Theorem 1] , we have the following transformation. Lemma 2.2 . We have \displaystyle\sum_{n\geq 0}\frac{\left(B,-Abq\right)_{n}q^{n}}{\left(-aq,-bq\right)_{n}} \displaystyle=-\frac{a^{-1}\left(B,-Abq\right)_{\infty}}{\left(-aq,-bq\right)_{\infty}}\sum_{n\geq 0}\frac{\left(A^{-1}\right)_{n}\left(\frac{Abq}{a}\right)^{n}}{\left(-\frac{B}{a}\right)_{n+1}} \displaystyle\hskip 71.13188pt+\left(1+b\right)\sum_{n\geq 0}\frac{\left(-a^{-1}\right)_{n+1}\left(-\frac{ABq}{a}\right)_{n}\left(-b\right)^{n}}{\left(-\frac{B}{a},\frac{Abq}{a}\right)_{n+1}}. We next state the Rogers–Fine identity [ 1 , equation (2.7)] . Lemma 2.3 . We have \sum_{n\geq 0}\frac{\left(\alpha\right)_{n}w^{n}}{\left(\beta\right)_{n}}=\sum_{n\geq 0}\frac{\left(\alpha,\frac{\alpha wq}{\beta}\right)_{n}\beta^{n}w^{n}q^{n^{2}-n}\left(1-\alpha wq^{2n}\right)}{\left(\beta\right)_{n}\left(w\right)_{n+1}}. The Gaussian polynomial is defined by (see [ 2 , Definition 3.1] ) \begin{bmatrix}n\\ m\end{bmatrix}_{q}:=\begin{cases}\frac{(q)_{n}}{(q)_{m}(q)_{n-m}}&\quad\text{if}\ 0\leq m\leq n,\\ 0&\quad\text{otherwise}.\end{cases} From [ 2 , equation (3.3.6)] , we have, for n\in\mathbb{N} , (2.1) \left(a;q\right)_{N}=\sum_{j=0}^{N}\begin{bmatrix}N\\ j\end{bmatrix}_{q}(-1)^{j}a^{j}q^{\frac{j^{2}-j}{2}}. 3. Some q -series transformations The following lemma is a key step for proving Theorem 1.2 . Lemma 3.1 . For k\in\mathbb{N}_{\geq 3} , we have F_{k,1}(q)=\frac{q}{\left(1-q\right)\left(q;q^{2}\right)_{k-1}}\left(1+\left(1-q\right)\sum_{n=1}^{k-2}\frac{\left(q^{4-2k};q^{2}\right)_{n}}{\left(q^{2};q^{2}\right)_{n}}\frac{q^{(2k-1)n}}{1-q^{2n+1}}\right). Proof. Using ( 1.1 ), we obtain (3.1) \displaystyle F_{k,1}(q) \displaystyle=\frac{q\left(q^{2};q^{2}\right)^{2}_{\infty}}{\left(q;q^{2}\right)^{2}_{\infty}\left(q^{2};q^{2}\right)_{k-1}}\sum_{n\geq 0}\frac{\left(q;q^{2}\right)^{2}_{n}q^{2n}}{\left(q^{2};q^{2}\right)_{n}\left(q^{2k};q^{2}\right)_{n}}. Applying Lemma 2.1 with q\mapsto q^{2} , a=b=q , c=q^{2k} , and t=q^{2} , we have \sum_{n\geq 0}\frac{\left(q;q^{2}\right)^{2}_{n}q^{2n}}{\left(q^{2},q^{2k};q^{2}\right)_{n}}=\frac{\left(q^{2k-1},q^{3};q^{2}\right)_{\infty}}{\left(q^{2k},q^{2};q^{2}\right)_{\infty}}\sum_{n\geq 0}\frac{\left(q^{4-2k},q;q^{2}\right)_{n}}{\left(q^{2},q^{3};q^{2}\right)_{n}}q^{(2k-1)n}. Plugging this into ( 3.1 ), it follows that (3.2) \displaystyle F_{k,1}(q) \displaystyle\!=\frac{q}{\left(1-q\right)\left(q;q^{2}\right)_{k-1}}\left(1+\left(1-q\right)\sum_{n\geq 1}\frac{\left(q^{4-2k};q^{2}\right)_{n}}{\left(q^{2};q^{2}\right)_{n}}\frac{q^{(2k-1)n}}{1-q^{2n+1}}\right). Note that for n\geq k-1 , (q^{4-2k};q^{2})_{n}=0 . Therefore, by ( 3.2 ), we have for k\geq 3 \displaystyle F_{k,1}(q)=\frac{q}{\left(1-q\right)\left(q;q^{2}\right)_{k-1}}\left(1+\left(1-q\right)\sum_{n=1}^{k-2}\frac{\left(q^{4-2k};q^{2}\right)_{n}}{\left(q^{2};q^{2}\right)_{n}}\frac{q^{(2k-1)n}}{1-q^{2n+1}}\right). This concludes the proof. ∎ Next, we present two lemmas which are used in the proof of Theorem 1.5 . Lemma 3.2 . For k\in\mathbb{N} , we have F_{k,1}(q)=\sum_{n\geq 0}\frac{\left(q^{2k-1};q^{2}\right)_{n}q^{n+1}}{\left(q;q^{2}\right)_{n+1}}. Proof. By ( 3.1 ), we have (3.3) F_{k,1}(q)=\frac{\left(q^{2};q^{2}\right)^{2}_{\infty}}{(q;q^{2})^{2}_{\infty}\left(q^{2};q^{2}\right)_{k-1}}\sum_{n\geq 0}\frac{\left(q;q^{2}\right)^{2}_{n}q^{2n+1}}{\left(q^{2},q^{2k};q^{2}\right)_{n}}. Taking q\mapsto q^{2} , a=-1,b=-q^{2k-2},A=q^{1-2k} , and B=q in Lemma 2.2 , using the fact that for n\in\mathbb{N}_{0} , (1;q^{2})_{n+1}=0 , we obtain \sum_{n\geq 0}\frac{\left(q;q^{2}\right)^{2}_{n}q^{2n+1}}{\left(q^{2},q^{2k};q^{2}\right)_{n}}=\frac{(q;q^{2})^{2}_{\infty}}{\left(q^{2},q^{2k};q^{2}\right)_{\infty}}\sum_{n\geq 0}\frac{\left(q^{2k-1};q^{2}\right)_{n}q^{n+1}}{\left(q;q^{2}\right)_{n+1}}. Plugging this into ( 3.3 ), we get \displaystyle F_{k,1}(q) \displaystyle=\frac{\left(q^{2};q^{2}\right)^{2}_{\infty}}{\left(q;q^{2}\right)^{2}_{\infty}\left(q^{2};q^{2}\right)_{k-1}}\frac{(q;q^{2})^{2}_{\infty}}{\left(q^{2},q^{2k};q^{2}\right)_{\infty}}\sum_{n\geq 0}\frac{\left(q^{2k-1};q^{2}\right)_{n}q^{n+1}}{\left(q;q^{2}\right)_{n+1}} \displaystyle=\sum_{n\geq 0}\frac{\left(q^{2k-1};q^{2}\right)_{n}q^{n+1}}{\left(q;q^{2}\right)_{n+1}}. This gives the lemma. ∎ 4. Proof of Theorems 1.2 , 1.3 , 1.4 , and 1.5 We are now ready to prove Theorem 1.2 . Proof of Theorem 1.2 . We show Theorem 1.2 in three parts. First we prove the claim for k=5 . By Lemma 3.1 , we have \displaystyle F_{5,1}(q) \displaystyle=\frac{q}{\left(1-q\right)\left(q;q^{2}\right)_{4}}\left(1+\left(1-q\right)\sum_{n=1}^{3}\frac{\left(q^{-6};q^{2}\right)_{n}}{\left(q^{2};q^{2}\right)_{n}}\frac{q^{9n}}{1-q^{2n+1}}\right) \displaystyle=\frac{q}{(1-q)\left(1-q^{3}\right)\left(1-q^{5}\right)\left(1-q^{7}\right)}\left(\frac{1}{1-q}+\frac{1-q^{-6}}{1-q^{2}}\frac{q^{9}}{1-q^{3}}\right. \displaystyle\hskip 14.22636pt\left.+\frac{\left(1-q^{-6}\right)\left(1-q^{-4}\right)}{\left(1-q^{2}\right)\left(1-q^{4}\right)}\frac{q^{18}}{1-q^{5}}+\frac{\left(1-q^{-6}\right)\left(1-q^{-4}\right)\left(1-q^{-2}\right)}{\left(1-q^{2}\right)\left(1-q^{4}\right)\left(1-q^{6}\right)}\frac{q^{27}}{1-q^{7}}\right). Simplifying the factors in the summands on the right-hand side, we obtain F_{5,1}(q)=\frac{q}{(1-q)\left(1-q^{3}\right)\left(1-q^{5}\right)\left(1-q^{7}\right)}\\ \times\left(\frac{1}{1-q}-\left(1+q^{2}+q^{4}\right)\frac{q^{3}}{1-q^{3}}+\left(1+q^{2}+q^{4}\right)\frac{q^{8}}{1-q^{5}}-\frac{q^{15}}{1-q^{7}}\right). The claim follows once we prove that \frac{1}{1-q}\left(\frac{1}{1-q}-\frac{\left(1+q^{2}+q^{4}\right)q^{3}}{1-q^{3}}+\frac{\left(1+q^{2}+q^{4}\right)q^{8}}{1-q^{5}}-\frac{q^{15}}{1-q^{7}}\right)\succeq 0. We now show that we can treat the first and the second together as well as the third and fourth. First we write \frac{\left(1+q^{2}+q^{4}\right)q^{3}}{1-q^{3}}=-1-q-q^{2}-q^{4}+\frac{1+q+q^{2}}{1-q^{3}}. Then \frac{1}{1-q}+1+q+q^{2}+q^{4}-\frac{1+q+q^{2}}{1-q^{3}}\succeq 0. Next we write \frac{\left(1+q^{2}+q^{4}\right)q^{8}}{1-q^{5}}-\frac{q^{15}}{1-q^{7}}=q^{8}\bigg(\frac{1+q^{2}+q^{4}}{1-q^{5}}+1-\frac{1}{1-q^{7}}\bigg). Now \frac{1+q^{2}+q^{4}}{1-q^{5}}-\frac{1}{1-q^{7}}\succeq\frac{1}{1-q^{5}}-\frac{1}{1-q^{7}}=\frac{q^{5}-q^{7}}{\left(1-q^{5}\right)\left(1-q^{7}\right)}=\frac{q^{5}\left(1-q^{2}\right)}{\left(1-q^{5}\right)\left(1-q^{7}\right)}. Dividing by \frac{1}{1-q} gives \frac{q^{5}(1+q)}{(1-q^{5})(1-q^{7})}\succeq 0. This finishes the proof of the case k=5 . Next, we show the case k=6 . By Lemma 3.1 , we have F_{6,1}(q)=\frac{q}{(1-q)(q;q^{2})_{5}}\left(1+(1-q)\sum_{n=1}^{4}\frac{\left(q^{-8};q^{2}\right)_{n}}{(q^{2};q^{2})_{n}}\frac{q^{11n}}{1-q^{2n+1}}\right). Simplifying the factors \frac{(q^{-8};q^{2})_{n}}{(q^{2};q^{2})_{n}} (for 1\leq n\leq 4 ) in the sum on the right-hand side of the above identity, we have F_{6,1}(q)=\frac{q}{(q;q^{2})_{5}}\left(\frac{1}{1-q}-\frac{q^{3}\left(1+q^{2}+q^{4}+q^{6}\right)}{1-q^{3}}+\frac{q^{8}\left(1+q^{2}+q^{4}\right)\left(1+q^{4}\right)}{1-q^{5}}\right.\\ \left.-\frac{q^{15}\left(1+q^{2}+q^{4}+q^{6}\right)}{1-q^{7}}+\frac{q^{24}}{1-q^{9}}\right). Using long division, we write \displaystyle-\frac{q^{3}\left(1+q^{2}+q^{4}+q^{6}\right)}{1-q^{3}}=2+q+q^{2}+q^{3}+q^{4}+q^{6}-\frac{2+q+q^{2}}{1-q^{3}}, \displaystyle\frac{q^{8}\left(1+q^{2}+q^{4}\right)\left(1+q^{4}\right)}{1-q^{5}}=-1-q-2q^{2}-q^{3}-q^{4}-q^{5}-q^{6}-2q^{7}-q^{9}-q^{11} \displaystyle\hskip 270.30118pt+\frac{1+q+2q^{2}+q^{3}+q^{4}}{1-q^{5}}, \displaystyle-\frac{q^{15}\left(1+q^{2}+q^{4}+q^{6}\right)}{1-q^{7}}=1+q+q^{3}+q^{5}+q^{7}+q^{8}+q^{10}+q^{12}+q^{14} \displaystyle\hskip 301.59938pt-\frac{1+q+q^{3}+q^{5}}{1-q^{7}}. Thus, we need \displaystyle\frac{1}{1-q}\left(\frac{1}{1-q}+2+q+q^{2}+q^{3}+q^{4}+q^{6}-\frac{2+q+q^{2}}{1-q^{3}}-1-q-2q^{2}-q^{3}-q^{4}\right. \displaystyle\left.-q^{5}-q^{6}-2q^{7}-q^{9}-q^{11}+\frac{1+q+2q^{2}+q^{3}+q^{4}}{1-q^{5}}+1+q+q^{3}+q^{5}+q^{7}+q^{8}\right. \displaystyle\hskip 210.55022pt\left.+q^{10}+q^{12}+q^{14}-\frac{1+q+q^{3}+q^{5}}{1-q^{7}}\right) \displaystyle=\frac{1}{1-q}\left(\frac{1}{1-q}+2+q-q^{2}+q^{3}-q^{7}+q^{8}-q^{9}+q^{10}-q^{11}+q^{12}+q^{14}\right. \displaystyle\hskip 81.09052pt\left.-\frac{2+q+q^{2}}{1-q^{3}}+\frac{1+q+2q^{2}+q^{3}+q^{4}}{1-q^{5}}-\frac{1+q+q^{3}+q^{5}}{1-q^{7}}\right)\succeq 0. Now we write \displaystyle\frac{1+q+2q^{2}+q^{3}+q^{4}}{1-q^{5}} \displaystyle=\frac{1}{1-q}+\frac{q^{2}}{1-q^{5}},\quad-\frac{2+q+q^{2}}{1-q^{3}}=-\frac{1}{1-q^{3}}-\frac{1}{1-q}, \displaystyle-\frac{1+q+q^{3}+q^{5}}{1-q^{7}} \displaystyle=-\frac{1}{1-q}+\frac{q^{2}+q^{4}+q^{6}}{1+q^{7}}. So we want \displaystyle\hskip-5.69046pt\frac{1}{1-q}\Biggl(2+q-q^{2}+q^{3}-q^{7}+q^{8}-q^{9}+q^{10}-q^{11}+q^{12}+q^{14}-\frac{1}{1-q^{3}} (4.1) \displaystyle\hskip 182.09746pt+\frac{q^{2}}{1-q^{5}}+\frac{q^{2}\left(1+q^{2}+q^{4}\right)}{1-q^{7}}\Biggr)\succeq 0. To show this, first note that \displaystyle\frac{1}{1-q}\!\left(-\frac{1}{1-q^{3}}\!+\!\frac{q^{2}}{1-q^{5}}\!+\!\frac{q^{2}\left(1+q^{2}+q^{4}\right)}{1-q^{7}}\right) (4.2) \displaystyle\hskip 147.95424pt\!\succeq\!\frac{1}{1-q}\!\left(\!-\frac{1}{1-q^{3}}\!+\!\frac{q^{2}}{1-q^{5}}\!+\!\frac{q^{2}\left(1+q^{2}\right)}{1-q^{7}}\right). Now we prove that (4.3) \frac{1}{1-q}\!\left(-\frac{1}{1-q^{3}}+\frac{q^{2}}{1-q^{5}}+\frac{q^{2}\left(1+q^{2}\right)}{1-q^{7}}\right)\succeq-\frac{1+q}{1-q}. This is equivalent to showing that (4.4) \frac{1}{1-q}\!\left(1+q-\frac{1}{1-q^{3}}+\frac{q^{2}}{1-q^{5}}+\frac{q^{2}\left(1+q^{2}\right)}{1-q^{7}}\right)\succeq 0. We have \displaystyle\hskip-28.45274pt\frac{1}{1-q}\!\left(1+q-\frac{1}{1-q^{3}}+\frac{q^{2}}{1-q^{5}}+\frac{q^{2}\left(1+q^{2}\right)}{1-q^{7}}\right) \displaystyle=\frac{1}{1-q}\!\left(1+q-\frac{1-q^{2}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{q^{2}\left(1+q^{2}\right)}{1-q^{7}}\right) \displaystyle=\frac{1+q}{1-q}-\frac{1+q}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{q^{2}+q^{4}}{\left(1-q\right)\left(1-q^{7}\right)} \displaystyle=\frac{1+q}{1-q}-\frac{1+q}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{q^{2}-q^{4}+2q^{4}}{\left(1-q\right)\left(1-q^{7}\right)} \displaystyle=\frac{1+q}{1-q}-\frac{1+q}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{q^{2}\left(1+q\right)}{1-q^{7}}+\frac{2q^{4}}{\left(1-q\right)\left(1-q^{7}\right)}. Since \frac{q^{2}\left(1+q\right)}{1-q^{7}}\succeq 0 , therefore, to prove ( 4.4 ), it is enough to show that (4.5) \frac{1+q}{1-q}-\frac{1+q}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{2q^{4}}{\left(1-q\right)\left(1-q^{7}\right)}\succeq 0. Simplifying the left-hand side of ( 4.5 ), we get \displaystyle\frac{1+q}{1-q}\!-\!\frac{1+q}{\left(1-q^{3}\right)\left(1-q^{5}\right)}\!+\!\frac{2q^{4}}{\left(1-q\right)\left(1-q^{7}\right)} \displaystyle\hskip 71.13188pt=\!\frac{\left(1+q\right)\left(q+q^{2}-q^{5}-q^{6}-q^{7}\right)}{\left(1-q^{3}\right)\left(1-q^{5}\right)}\!+\!\frac{2q^{4}}{\left(1-q\right)\left(1-q^{7}\right)} \displaystyle\hskip 71.13188pt=\frac{\left(1+q\right)\left(q\left(1-q^{6}\right)+q^{2}\left(1-q^{3}\right)-q^{6}\right)}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{2q^{4}}{\left(1-q\right)\left(1-q^{7}\right)} \displaystyle\hskip 71.13188pt\succeq q^{4}\left(-\frac{q^{2}+q^{3}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}+\frac{2}{\left(1-q\right)\left(1-q^{7}\right)}\right) \displaystyle\hskip 71.13188pt=q^{4}\Bigg(\frac{2q}{\left(1-q^{3}\right)\left(1-q^{7}\right)}+\frac{2\left(1+q^{2}\right)}{\left(1-q^{3}\right)\left(1-q^{7}\right)}-\frac{q^{2}+q^{3}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}\Bigg). Since \frac{2q}{\left(1-q^{3}\right)\left(1-q^{7}\right)}\succeq 0 , thus, to prove ( 4.5 ), it suffices to show that \frac{2\left(1+q^{2}\right)}{\left(1-q^{3}\right)\left(1-q^{7}\right)}-\frac{q^{2}+q^{3}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}\succeq 0. To prove this, we show that (4.6) \displaystyle\frac{1+q^{2}}{\left(1-q^{3}\right)\left(1-q^{7}\right)}-\frac{q^{2}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}\succeq 0, (4.7) \displaystyle\frac{1+q^{2}}{\left(1-q^{3}\right)\left(1-q^{7}\right)}-\frac{q^{3}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}\succeq 0. First, we see that \displaystyle\frac{1+q^{2}}{\left(1-q^{3}\right)\left(1-q^{7}\right)}-\frac{q^{2}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}=\frac{1}{1-q^{3}}+\frac{q^{9}}{\left(1-q^{5}\right)\left(1-q^{7}\right)}\succeq 0, which proves ( 4.6 ). Next, we have \displaystyle\frac{1+q^{2}}{\left(1-q^{3}\right)\left(1-q^{7}\right)}-\frac{q^{3}}{\left(1-q^{3}\right)\left(1-q^{5}\right)}=\frac{1}{\left(1-q^{5}\right)\left(1-q^{7}\right)}+\frac{q^{2}}{1-q^{7}}\succeq 0. This gives ( 4.7 ). This concludes the proof of ( 4.3 ). Thus, combining ( 4 ) and ( 4.3 ), it follows that \frac{1}{1-q}\!\left(-\frac{1}{1-q^{3}}\!+\!\frac{q^{2}}{1-q^{5}}\!+\!\frac{q^{2}\left(1+q^{2}+q^{4}\right)}{1-q^{7}}\right)\!\succeq-\frac{1+q}{1-q}. Thus, to prove ( 4 ), it remains to show that \frac{1}{1-q}\left(2+q-q^{2}+q^{3}-q^{7}+q^{8}-q^{9}+q^{10}-q^{11}+q^{12}+q^{14}-\left(1+q\right)\right)\succeq 0, which holds. This gives the claim for k=6 . Finally, we consider the case k=7 . From Lemma 3.1 , we have F_{7,1}(q)=\frac{q}{\left(1-q\right)\left(q;q^{2}\right)_{6}}\left(1+(1-q)\sum_{n=1}^{5}\frac{\left(q^{-10};q^{2}\right)_{n}}{\left(q^{2};q^{2}\right)_{n}}\frac{q^{13n}}{1-q^{2n+1}}\right). Simplifying the factors \frac{(q^{-10};q^{2})_{n}}{(q^{2};q^{2})_{n}} (for 1\leq n\leq 5 ) in the sum on the right hand side of the above identity, we have \displaystyle F_{7,1}(q) \displaystyle=\frac{q}{\left(q;q^{2}\right)_{6}}\left(\frac{1}{1-q}-\frac{q^{3}\left(1+q^{2}+q^{4}+q^{6}+q^{8}\right)}{1-q^{3}}\right. \displaystyle\hskip-28.45274pt\quad\,\left.+\frac{q^{8}\left(1+q^{4}\right)\left(1+q^{2}+q^{4}+q^{6}+q^{8}\right)}{1-q^{5}}-\frac{q^{15}\left(1+q^{4}\right)\left(1+q^{2}+q^{4}+q^{6}+q^{8}\right)}{1-q^{7}}\right. \displaystyle\hskip 133.72786pt\quad\,\left.+\frac{q^{20}\left(1+q^{2}+q^{4}+q^{6}+q^{8}\right)}{1-q^{9}}-\frac{q^{35}}{1-q^{11}}\right). Similar to the case k=6 , after carrying long divisions, to prove the claim, we need to show that \displaystyle\frac{1}{\left(q;q^{2}\right)_{6}}\Biggl(2-q+2q^{2}-q^{4}+q^{5}-q^{6}-q^{10}-2q^{11}+2q^{12}+2q^{14}-2q^{15} \displaystyle\hskip 85.35826pt+2q^{16}-q^{17}+q^{18}-q^{19}+q^{20}+q^{24}+\frac{1}{1-q}-\frac{2+q+2q^{2}}{1-q^{3}} \displaystyle\hskip 85.35826pt+\frac{2+2q+2q^{2}+2q^{3}+2q^{4}}{1-q^{5}}-\frac{2+q+2q^{2}+q^{3}+q^{4}+2q^{5}+q^{6}}{1-q^{7}} \displaystyle\hskip 184.9429pt+\frac{q+q^{2}+q^{4}+q^{6}+q^{8}}{1-q^{9}}-\frac{q^{2}}{1-q^{11}}\Biggr)\succeq 0. Now we write \displaystyle-\frac{2+q+2q^{2}}{1-q^{3}} \displaystyle=-\frac{1}{1-q}-\frac{1+q^{2}}{1-q^{3}}, \displaystyle\frac{2+2q+2q^{2}+2q^{3}+2q^{4}}{1-q^{5}} \displaystyle=\frac{2}{1-q}, \displaystyle-\frac{2+q+2q^{2}+q^{3}+q^{4}+2q^{5}+q^{6}}{1-q^{7}} \displaystyle=-\frac{1}{1-q}-\frac{1+q^{2}+q^{5}}{1-q^{7}}, \displaystyle\frac{q+q^{2}+q^{4}+q^{6}+q^{8}}{1-q^{9}} \displaystyle=\frac{1}{1-q}-\frac{1+q^{3}+q^{5}+q^{7}}{1-q^{9}}. Thus, note that the claim follows if we show \displaystyle\frac{1}{1-q}\Biggl(2-q+2q^{2}-q^{4}+q^{5}-q^{6}-q^{10}-2q^{11}+2q^{12}+2q^{14}-2q^{15}+2q^{16} \displaystyle\hskip 14.22636pt-\!q^{17}\!+\!q^{18}\!-\!q^{19}\!-\!\frac{1\!+\!q^{2}}{1\!-\!q^{3}}\!+\!\frac{2}{1\!-\!q}\!-\!\frac{1\!+\!q^{2}\!+\!q^{5}}{1\!-\!q^{7}}\!-\!\frac{1\!+\!q^{3}\!+\!q^{5}\!+\!q^{7}}{1\!-\!q^{9}}\!-\!\frac{q^{2}}{1\!-\!q^{11}}\Biggr)\succeq 0. Next, we simplify \displaystyle-\frac{1+q^{2}}{1-q^{3}}+\frac{2}{1-q}-\frac{1+q^{3}+q^{5}+q^{7}}{1-q^{9}} \displaystyle=\frac{2q+q^{2}+2q^{4}+q^{6}+q^{7}+q^{8}}{1-q^{9}}. Thus, we need to show that \displaystyle\frac{1}{1-q}\Biggl(2-q+2q^{2}-q^{4}+q^{5}-q^{6}-q^{10}-2q^{11}+2q^{12}+2q^{14}-2q^{15}+2q^{16} (4.8) \displaystyle\hskip 0.0pt\!-\!q^{17}\!+\!q^{18}\!-\!q^{19}\!+\!\frac{2q\!+\!q^{2}\!+\!2q^{4}\!+\!q^{6}\!+\!q^{7}\!+\!q^{8}}{1\!-\!q^{9}}\!-\!\frac{1\!+\!q^{2}\!+\!q^{5}}{1\!-\!q^{7}}\!-\!\frac{q^{2}}{1\!-\!q^{11}}\Biggr)\succeq 0. Note that \displaystyle\frac{2\!-\!q\!+\!2q^{2}\!-\!q^{4}\!+\!q^{5}\!-\!q^{6}\!-\!q^{10}\!-\!2q^{11}\!+\!2q^{12}\!+\!2q^{14}\!-\!2q^{15}\!+\!2q^{16}\!-\!q^{17}\!+\!q^{18}\!-\!q^{19}}{1-q} \displaystyle=\frac{1-2q^{11}}{1-q}+\frac{q^{2}\left(1-q^{8}\right)}{1-q}+\frac{2q^{12}+q^{16}}{1-q}+1+q^{2}+q^{3}+q^{5}+2q^{14}+q^{16}+q^{18}. Since \frac{q^{2}\left(1-q^{8}\right)}{1-q}+\frac{2q^{12}+q^{16}}{1-q}+1+q^{2}+q^{3}+q^{5}+2q^{14}+q^{16}+q^{18}\succeq 0, to prove ( 4 ), it suffices to show \displaystyle\frac{1}{1-q}\left(1-2q^{11}+\frac{2q+q^{2}+2q^{4}+q^{6}}{1-q^{9}}-\frac{1+q^{2}+q^{5}}{1-q^{7}}-\frac{q^{2}}{1-q^{11}}\right)\succeq 0. To prove this, we prove the following (4.9) \displaystyle\frac{1}{1-q}\left(1-2q^{11}+\frac{q+q^{2}+2q^{4}+q^{6}}{1-q^{9}}-\frac{1+q^{2}+q^{5}}{1-q^{7}}\right) \displaystyle\succeq 0, (4.10) \displaystyle\frac{1}{1-q}\left(\frac{q}{1-q^{9}}-\frac{q^{2}}{1-q^{11}}\right) \displaystyle\succeq 0. Note that \frac{1}{1-q}\!\left(\frac{q}{1-q^{9}}\!-\!\frac{q^{2}}{1-q^{11}}\right)\!=\!\frac{q\!-\!q^{12}\!-\!q^{2}\!+\!q^{11}}{\left(1\!-\!q\right)\left(1\!-\!q^{9}\right)\left(1\!-\!q^{11}\right)}\!=\!\frac{q\!+\!q^{11}}{\left(1\!-\!q^{9}\right)\left(1\!-\!q^{11}\right)}\succeq 0, which proves ( 4.10 ). Next we prove ( 4.9 ). First we see that \displaystyle\frac{1}{1-q}\left(1-2q^{11}+\frac{q+q^{2}+2q^{4}+q^{6}}{1-q^{9}}-\frac{1+q^{2}+q^{5}}{1-q^{7}}\right) \displaystyle\hskip 28.45274pt\succeq 1+\frac{1}{1-q}\left(\frac{q+q^{2}+2q^{4}+q^{6}-q^{11}+q^{20}}{1-q^{9}}-\frac{1+q^{2}+q^{5}}{1-q^{7}}\right) \displaystyle\hskip 28.45274pt=1+\frac{1}{1-q}\left(\frac{q+2q^{4}+q^{6}+q^{20}}{1-q^{9}}-\frac{1+q^{5}+q^{9}}{1-q^{7}}\right) \displaystyle\hskip 28.45274pt\succeq 1+\frac{1}{1-q}\left(\frac{q+2q^{4}+q^{6}}{1-q^{9}}-\frac{1+q^{5}+q^{9}}{1-q^{7}}\right) \displaystyle\hskip 28.45274pt=\frac{q^{4}\left(1+q+q^{2}+q^{3}+q^{4}\right)}{1-q^{9}}+\frac{1}{1-q}\left(\frac{1+q^{4}+q^{6}+q^{10}}{1-q^{9}}-\frac{1+q^{5}+q^{9}}{1-q^{7}}\right). Since \frac{q^{4}\left(1+q+q^{2}+q^{3}+q^{4}\right)}{1-q^{9}}\succeq 0, to prove ( 4.9 ), it is enough to show that \frac{1}{1-q}\left(\frac{1+q^{4}+q^{6}+q^{10}}{1-q^{9}}-\frac{1+q^{5}+q^{9}}{1-q^{7}}\right)\succeq 0. We write \frac{q^{10}}{1-q^{9}}=-q+\frac{q}{1-q^{9}},\quad\frac{q^{9}}{1-q^{7}}=-q^{2}+\frac{q^{2}}{1-q^{7}}. Thus we need to show that \frac{1}{1-q}\left(-q+q^{2}+\frac{1+q+q^{4}+q^{6}}{1-q^{9}}-\frac{1+q^{2}+q^{5}}{1-q^{7}}\right)\succeq 0. We combine \displaystyle\frac{1\!+\!q\!+\!q^{4}\!+\!q^{6}}{1\!-\!q^{9}}\!-\!\frac{1\!+\!q^{2}\!+\!q^{5}}{1\!-\!q^{7}} \displaystyle=\frac{\left(1\!-\!q\right)q\left(1\!+\!q^{3}\!+\!q^{5}\!-\!q^{7}\!-\!q^{12}\right)}{\left(1\!-\!q^{7}\right)\left(1\!-\!q^{9}\right)}. So we need -1+\frac{1+q^{3}+q^{5}-q^{7}-q^{12}}{\left(1-q^{7}\right)\left(1-q^{9}\right)}\succeq 0. It is not hard to see that the constant term of the left-hand side vanishes. Thus it is enough to show that \frac{1+q^{3}-q^{7}-q^{12}}{\left(1-q^{7}\right)\left(1-q^{9}\right)}=\frac{1}{1-q^{9}}+\frac{q^{3}}{1-q^{7}}\succeq 0 which holds. thus concludes the proof for k=7 . ∎ . Next, we prove Theorem 1.3 . Proof of Theorem 1.3 . Throughout the proof, k\!\in\!\{8,9,10\} . We begin by establishing an alternative form for F_{k,1}(q) . To make the powers in the inner sum of Lemma 3.1 more explicitly positive, we distribute powers of q . For this, we write \displaystyle\left(q^{4-2k};q^{2}\right)_{n}q^{(2k-1)n}=(-1)^{n}\left(q^{2k-2n-2};q^{2}\right)_{n}q^{n^{2}+2n}. Plugging this into Lemma 3.1 , we obtain \displaystyle F_{k,1}(q) \displaystyle=\frac{q}{(1-q)(q;q^{2})_{k-1}}\left(1+(1-q)\sum_{n=1}^{k-2}\frac{(-1)^{n}\left(q^{2k-2n-2};q^{2}\right)_{n}q^{n^{2}+2n}}{(1-q^{2n+1})(q^{2};q^{2})_{n}}\right) \displaystyle=\frac{1}{(q;q^{2})_{k-1}}\!\sum_{n=0}^{k-2}\!\frac{(-1)^{n}\left(q^{2k-2n-2};q^{2}\right)_{n}q^{(n+1)^{2}}}{(1-q^{2n+1})(q^{2};q^{2})_{n}} \displaystyle=\frac{1}{(q;q^{2})_{k-1}}\!\sum_{n=0}^{k-2}\!\begin{bmatrix}k-2\\ n\end{bmatrix}_{q^{2}}\!\!\!\frac{(-1)^{n}q^{(n+1)^{2}}}{1\!-\!q^{2n+1}}. Note that we can place all terms in the inner sum over the common denominator 1-q^{\ell} , where \ell:=\text{lcm}(1,3,5,\dots,2k-3) . Thus we obtain F_{k,1}(q)=\frac{1}{(q;q^{2})_{k-1}(1-q^{\ell})}\sum_{n=0}^{k-2}\begin{bmatrix}k-2\\ n\end{bmatrix}_{q^{2}}(-1)^{n}q^{(n+1)^{2}}\sum_{j=0}^{\frac{\ell}{2n+1}-1}q^{\left(2n+1\right)j}. Now F_{k,1}(q)\succeq 0 holds if (4.11) \frac{1}{1-q}\sum_{n=0}^{k-2}\begin{bmatrix}k-2\\ n\end{bmatrix}_{q^{2}}(-1)^{n}q^{(n+1)^{2}}\sum_{j=0}^{\frac{\ell}{2n+1}-1}q^{\left(2n+1\right)j}\succeq 0. Now consider the inner sum, which is the polynomial H_{k}(q):=\sum_{n=0}^{k-2}\begin{bmatrix}k-2\\ n\end{bmatrix}_{q^{2}}(-1)^{n}q^{(n+1)^{2}}\sum_{j=0}^{\frac{\ell}{2n+1}-1}q^{\left(2n+1\right)j}=:\sum_{j=0}^{d_{k}}a_{k}(j)q^{j}, where d_{k} is the degree of H_{k}(q) , and a_{k}(j):=\text{coeff}_{[q^{j}]}(H_{k}(q)) . For a polynomial \mathcal{P}(q)=\sum_{j=0}^{d}a(j)q^{j} of degree d ( d\in\mathbb{N} ) with a(j)\in\mathbb{R} and a(j)=0 for j>d , we write \displaystyle\frac{\mathcal{P}(q)}{1-q}=\sum_{m\geq 0}\sum_{j=0}^{d}a(j)q^{j+m}=\sum_{n\geq 0}c(n)q^{n}, using j+m=n and a(n)>0 for n>d in the final step, where c(n):=\begin{cases}\sum_{j=0}^{n}a(j)\ \ \text{if}\ \ 0\leq n\leq d,\\ \sum_{j=0}^{d}a(j)\ \ \text{if}\ \ n>d.\end{cases} Therefore, to prove \frac{\mathcal{P}(q)}{1-q}\succeq 0 , it suffices to check that c(n)\geq 0 for 0\leq n\leq d . Using this with \mathcal{P}(q)=H_{k}(q) , we see that to show ( 4.11 ), it is enough to check that \sum_{j=0}^{m}a_{k}(j)\geq 0 for all 0\leq m\leq d_{k} . We have checked this with a computer and the code 1 1 1 We provide the code for k=10 . The code for k\!\in\!\{8,9\} is similar. is provided in Appendix A.∎ Next, we prove Theorem 1.4 . Proof of Theorem 1.4 . By Lemma 3.2 , it follows that (4.12) \lim_{k\to\infty}F_{k,1}(q)=q\sum_{n\geq 0}\frac{q^{n}}{\left(q;q^{2}\right)_{n+1}}. Applying Lemma 2.3 with q\mapsto q^{2} , \alpha=0 , \beta=q^{3} , and w=q , and then multiplying both sides by \frac{q}{1-q} and finally using ( 1.2 ), we obtain (4.13) q\sum_{n\geq 0}\frac{q^{n}}{\left(q;q^{2}\right)_{n+1}}=q\omega(q). Plugging this into ( 4.12 ), we conclude the proof. ∎ We are now ready to prove Theorem 1.5 . Proof of Theorem 1.5 . By ( 4.13 ) and Lemma 3.2 , we have \displaystyle q\omega(q)-F_{k,1}(q) \displaystyle=\sum_{n\geq 0}\frac{q^{n+1}}{\left(q;q^{2}\right)_{n+1}}-\sum_{n\geq 0}\frac{\left(q^{2k-1};q^{2}\right)_{n}q^{n+1}}{\left(q;q^{2}\right)_{n+1}} (4.14) \displaystyle=\sum_{n\geq 0}\frac{1-\left(q^{2k-1};q^{2}\right)_{n}}{\left(q;q^{2}\right)_{n+1}}q^{n+1}. Applying ( 2.1 ) with q\mapsto q^{2},a=q^{2k-1} , and N=n , it follows that \left(q^{2k-1};q^{2}\right)_{n}=\sum_{j=0}^{n}\begin{bmatrix}n\\ j\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj-2j}. Consequently, we have 1-\left(q^{2k-1};q^{2}\right)_{n}=1-\sum_{j=0}^{n}\begin{bmatrix}n\\ j\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj-2j}=-\sum_{j=1}^{n}\begin{bmatrix}n\\ j\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj-2j}. Plugging this into ( 4 ), we get \displaystyle q\omega(q)-F_{k,1}(q)=\sum_{n\geq 0}\frac{q^{n+1}}{\left(q;q^{2}\right)_{n+1}}\sum_{j=1}^{n}\begin{bmatrix}n\\ j\end{bmatrix}_{q^{2}}(-1)^{j+1}q^{j^{2}+2kj-2j} \displaystyle=\sum_{n\geq 0}\frac{q^{n}}{\left(q;q^{2}\right)_{n+1}}\sum_{j=1}^{n}\begin{bmatrix}n\\ j\end{bmatrix}_{q^{2}}(-1)^{j+1}q^{\left(j-1\right)^{2}+2kj} \displaystyle=\sum_{n\geq 1}\frac{q^{n}}{\left(q;q^{2}\right)_{n+1}}\sum_{j=0}^{n-1}\begin{bmatrix}n\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj+2k} \displaystyle=q^{2k}\left(\frac{q}{\left(1-q\right)\left(1-q^{3}\right)}+\sum_{n\geq 2}\frac{q^{n}}{\left(q;q^{2}\right)_{n+1}}\sum_{j=0}^{n-1}\begin{bmatrix}n\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj}\right) \displaystyle=q^{2k+1}\left(\frac{1}{\left(1-q\right)\left(1-q^{3}\right)}+\sum_{n\geq 2}\frac{q^{n-1}}{\left(q;q^{2}\right)_{n+1}}\sum_{j=0}^{n-1}\begin{bmatrix}n\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj}\right) \displaystyle=q^{2k+1}\left(\frac{1}{(1-q)(1-q^{3})}\right. \displaystyle\hskip 96.73918pt\left.+\frac{1}{(1-q)(1-q^{3})}\sum_{n\geq 1}\frac{q^{n}}{(q^{5};q^{2})_{n}}\sum_{j=0}^{n}\begin{bmatrix}n+1\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj}\right) \displaystyle=\frac{q^{2k+1}}{\left(1-q\right)\left(1-q^{3}\right)}\left(1+\sum_{n\geq 1}\frac{q^{n}}{\left(q^{5};q^{2}\right)_{n}}\sum_{j=0}^{n}\begin{bmatrix}n+1\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj}\right) \displaystyle=q^{2k+1}E_{k}(q), where E_{k}(q):=\frac{1}{\left(1-q\right)\left(1-q^{3}\right)}\left(1+\sum_{n\geq 1}\frac{q^{n}}{\left(q^{5};q^{2}\right)_{n}}\sum_{j=0}^{n}\begin{bmatrix}n+1\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj}\right). Next note that E_{k}(q)\in\mathbb{Z}[[q]], because \frac{1}{\left(1-q\right)\left(1-q^{3}\right)}\in\mathbb{Z}[[q]]\ \ \text{and}\ \ \begin{bmatrix}n+1\\ j+1\end{bmatrix}_{q^{2}}\in\mathbb{Z}[q]. Moreover, we have \operatorname{coeff}_{\left[q^{0}\right]}\left(\frac{1}{\left(1-q\right)\left(1-q^{3}\right)}\right)=1, and \operatorname{coeff}_{\left[q^{0}\right]}\left(\sum_{n\geq 1}\frac{q^{n}}{\left(q^{5};q^{2}\right)_{n}}\sum_{j=0}^{n}\begin{bmatrix}n+1\\ j+1\end{bmatrix}_{q^{2}}(-1)^{j}q^{j^{2}+2kj}\right)=0, because for n,k\in\mathbb{N} , 0\leq j\leq n , j^{2}+2kj+n\neq 0 . Consequently, \operatorname{coeff}_{\left[q^{0}\right]}E_{k}(q)=1. This concludes the proof.∎ 5. Potential extensions of current approaches Each of the approaches employed in this paper runs into difficulties, which we now discuss, followed by some suggestions on extensions or alternatives. In the proof of Theorem 1.2 , for k=5 , we first simplify [\begin{smallmatrix}k-2\\ n\end{smallmatrix}]_{q^{2}} , then do long division and combine. However we are unable to find a uniform process of simplification, as shown in the proof of Theorem 1.2 for k\in\{6,7\} . If a persistent pattern could be found then the entire conjecture might be provable in this way. In the proof of Theorem 1.3 , the calculation is algorithmic but the lcm grows exponentially in k . For k=11 we are unable to continue. A more careful choice of where to cancel might allow further calculation; sufficient systematic insight, possibly including some combinatorial injections, might even be able to fully establish the conjecture. We make the following connection with q -binomial quotients, which may hold independent interest for some readers. Recall Lemma 3.2 (5.1) F_{k,1}(q)=\sum_{n\geq 0}\frac{\left(q^{2k-1};q^{2}\right)_{n}q^{n+1}}{\left(q;q^{2}\right)_{n+1}}=:\sum_{n\geq 0}G_{k,n}(q)q^{n+1}. We make the following refined conjecture, which would in turn imply 1.1 . Next, we rewrite G_{k,n}(q) as a quotient of q -binomials. (5.2) G_{k,n}(q)=\frac{1}{1-q^{2n+1}}\frac{\begin{bmatrix}2k+2n-2\\ 2n\end{bmatrix}_{q}}{\begin{bmatrix}k+n-1\\ n\end{bmatrix}_{q^{2}}}. Conjecture 5.1 . We have G_{k,n}(q)\succeq 0 . The form given in ( 5.2 ), the symmetry of q -binomial coefficients about their central peaks, and computational data suggest the following further conjecture: Conjecture 5.2 . For n\in\mathbb{N} , k\geq n+1 , we have G_{k,n}-G_{k+1,n-1}\succeq 0. We can establish the following. Lemma 5.3 . For n\in\mathbb{N} , 1\leq j\leq n , G_{n+2-j,n}=G_{n+2,n-j} . Proof. Using ( 5.1 ) with k=n-j+2 and then using ( 5.1 ) with k=n+2 and n\mapsto n-j , we conclude the proof. ∎ Since G_{k,0}(q)=\frac{1}{1-q} has nonnegative coefficients for all k , it follows from Lemma 5.3 that Conjecture 5.2 would imply the nonnegativity of all G_{k,n} and hence 1.1 . Lemma 5.3 also has the consequence that if we can show that G_{k,n}\succeq 0 for each k and 0\leq n\leq k-2 , then all G_{k,n}\succeq 0 and Conjecture 1.1 would be proved. Finally, the forms exhibited in the proof of Theorem 1.2 , along with computational evidence, further suggest the following conjecture, which would also be sufficient to establish the positivity of F_{k,1}(q) : Conjecture 5.4 . We have G_{k,n}(q)=\sum_{j=0}^{n}\frac{z_{k,j}(q)}{1-q^{2j+1}} where all z_{k,j}(q) have nonnegative coefficients. Appendix A Code We include in this section the Maple code for the case k=10 that verifies Theorem 1.3 . . > eightpolys:=1, 1+q^2+q^4+q^6+q^8+q^10+q^12+q^14, 1+q^2+2q^4 +2q^6+3q^8+3q^10+4q^12+3q^14+3q^16+2q^18+2q^20+q^22+q^24, 1+q^2+2q^4+3q^6+4q^8+5q^10+6q^12+6q^14+6q^16+6q^18+5q^20+4q^22 +3q^24+2q^26+q^28+q^30, 1+q^2+2q^4+3q^6+5q^8+5q^10+7q^12+7q^14 +8q^16+7q^18+7q^20+5q^22+5q^24+3q^26+2q^28+q^30+q^32, 1+q^2 +2q^4+3q^6+4q^8+5q^10+6q^12+6q^14+6q^16+6q^18+5q^20+4q^22+3q^24 +2q^26+q^28+q^30, 1+q^2+2q^4+2q^6+3q^8+3q^10+4q^12+3q^14+3q^16 +2q^18+2q^20+q^22+q^24, 1+q^2+q^4+q^6+q^8+q^10+q^12+q^14, 1 > tenbigpoly := (eightpolys[1]) (q^((0+1)^2))((-1)^0) add(q^(j(2(0) + 1)), j=0..((765765/(2(0)+1))-1)) + (eightpolys[1+1]) (q^((1+1)^2))((-1)^1) add(q^(j(2(1) + 1)), j=0..((765765/(2(1)+1))-1)) + (eightpolys[2+1]) (q^((2+1)^2))((-1)^2) add(q^(j(2(2) + 1)), j=0..((765765/(2(2)+1))-1)) + (eightpolys[3+1]) (q^((3+1)^2))((-1)^3) add(q^(j(2(3) + 1)), j=0..((765765/(2(3)+1))-1)) + (eightpolys[4+1]) (q^((4+1)^2))((-1)^4) add(q^(j(2(4) + 1)), j=0..((765765/(2(4)+1))-1)) + (eightpolys[5+1]) (q^((5+1)^2))((-1)^5) add(q^(j(2(5) + 1)), j=0..((765765/(2(5)+1))-1)) + (eightpolys[6+1]) (q^((6+1)^2))((-1)^6) add(q^(j(2(6) + 1)), j=0..((765765/(2(6)+1))-1)) + (eightpolys[7+1]) (q^((7+1)^2))((-1)^7) add(q^(j(2(7) + 1)), j=0..((765765/(2(7)+1))-1)) + (eightpolys[8+1]) (q^((8+1)^2))((-1)^8) add(q^(j(2(8) + 1)), j=0..((765765/(2(8)+1))-1)): > currsum :=0; errorfound:=False; for s from 1 to 765830 do currsum := currsum + coeff(tenbigpoly, q^s): if currsum < 0 then print("Error found at", s); errorfound := True; end if; end do: if errorfound = False then print("No error found") end if .