In Puzzle #1, you have no good options to keep a good leave while scoring a reasonable number of points using the Z in the process. F6 ZO(O)N for 33 points scores the most number of points, it retains the very-weak BNUV combination in the future turn. As a result, it is precious little else to draw into any possible seven- and eight-letter bingos (SUBVENE(-d/s, which is the only possible seven-letter bingo that can be drawn into the BNUV leave, with the EES draw, and is likely be elusive if the S hook onto ZOON was blocked), as well as UNVIABLE, OVERBURN, and VIBURNUM) since they require some extremely specific tile draws from the bag that include only five out of 21 different consonants (D, L, M, R, and S), and there would be far more instances when you are saddled with a low-scoring liability (scoring only in the 10-20 point range on average, or with the lower 20-30 point range that keep really bad rack leaves) in the future turns with very few options to keep a good leave in the subsequent turns or potentially force you to exchange with varieties of bad three-tile draws that don't involve the S or the blanks, such as draws containing both of either UU/UV/UW plus one additional clunker aside from the G (which pairs very well with the N for potential -ING words), three-vowel draws that don't contain an E (such as the IOU draw), three-consonant draws in which all of them are incompatible together, draws that involve duplicated vowels other than duplicated U's, and draws containing one U combined with two other consonants that are incompatible together. Even if the Q turns up as you keep the U, it's almost nowhere to use the QU combination effectively for any reasonable number of points and you have to retain it for another turn to realize the future potential despite the QU combination being normally bingo-unfriendly.

The latter two options are better despite their point sacrifices. I7 Z(E)BU and I4 B(O)NZE both at least maximize some outside bingo chances by increasing the entropy of the two rack leaves due to the more flexible word constructions. I7 Z(E)BU places the awkward U next to the triple-letter score, which is not going to easy to make potentially high-scoring parallel plays, since the only two consonants that make two-letter words on the right of the U are the H and the M, both are also outstanding for bingoing due to the large word formation role. If your opponent has either the H or the M, but without any bingos or palatable options elsewhere on the board, your opponent is likely to create some counterplays that open some potentials. I4 B(O)NZE blows up the rack and draw one extra tile to potentially snag some good tiles to save the normally awkward NUV leave, especially the E's, L's, R's, S's, and the blanks, and maximizes some bingo possibilities into some freakish range, as it at least breaks up part of the very-awkward BUV mix, but it opens the potentially lucrative back R/S hooks onto BONZE that is unlikely to open the new S hook to your opponent to use it while giving completely different counterplays that might not help with the NUV leave.

Thus, you better off playing ZEBU and withhold some extra bingo possibilities at the cost of five points and keeping duplicated N's: the NNOV leave after ZEBU can be drawn into some seven-letter words that are likely parts of everyday English, such as CONNIVE, CONVENE, CONVENT, ENVIRON, NOVENAE/NOVENAS, INWOVEN and UNWOVEN with the W draw, as well as a handful of eight-letter bingos such as CONVINCE, COVENANT, ENVISION, INNOVATE, INVASION, and INVENTOR. Quackle simulation (6-ply) can't separate the winning chances between ZOON and ZEBU, but the latter play will allow you to score over 3 points more on average and bingoing at roughly 6% at a time with the NNOV leave, whereas the BNUV leave after ZOON bingos at only a measly 1% at a time, which is much worse, making ZEBU a much stronger play as a part of counterintuitive thinking. It is important to note that anti-synergies are much bigger deal than duplicate letters in general, especially with comparison using consonants that duplicate reasonably well versus the V's horrible synergy with most high-pointers as well as the U and the two-point G. In fact, while the BV combination itself is anti-synergistic, it combines very well with one or both of L/R as the B and the V on their own have some good synergies with the L and the R, but BV goes horribly with the N and the T due to the latter two letters' weaker synergies with the B and the V on their own. In fact, there are 3,144 seven- and eight-letter words that can be made with duplicated N's, often found in -ING words, whereas the BV combination makes up only a measly of 280 words, making the BNUV combination an extremely inflexible combination with a vast majority of three-tile draws.

As it turns out, Quackle simulation errantly prefers ZOON narrowly over ZEBU due to multiple litany of things that the simulation could get wrong, perhaps due to that Quackle miscalculates and underestimates the differences between the BNUV and NNOV leaves due to the specific entropy in terms of bingoing and scoring, and defense is not particularly worth it at an early deficit. First of all, Quackle would try to grab the extra 5 points on ZOON over ZEBU, leaving the horrific BNUV leave while neglecting the other factors regarding investing in the future, such as attempting to maximize bingo chances and increase the ability to wade through tiles that would both score a reasonable number of points and keep a reasonable leave. Even though the latter play would give your opponent slightly better average score, it actually divert your opponent to inadvertently provide multiple additional floaters for potential eight-letter bingos (especially if the C turns up). Quackle also overlooks the fact that the S-hook possibility for SUBVENE will often be blocked if your opponent has at least one S on their own that you don't have it. This is because Quackle is unaware of keeping horrific rack leaves with minimal scoring potentials or bingo possibilities after hundreds of simulation iterations.

It is also worth noting that if you're a beginner or an intermediate player, you should better off playing ZEBU over ZOON due to that the BNUV leave is much harder to manage in the future turns than the NNOV leave, as there are very few words that contain all of BNUV in them as opposed to NNOV, which the latter one at least permit some everyday English words to be played.