I read GuySrinivasan's comment about offensive and defensive spells, everyone picking at least one offensive and at least on defensive, and their mana affinities being distributed fairly, and confirm this with my own observations. 

Checking how individual spells vary based on individual mana strengths (todo - how do they vary with combinations of strengths), indeed each one varies strongest with the two affinities you'd expect. Also, I noticed that if you form a chain of offensive spells based on common affinities, you form a single loop. The defensive spells on the other hand form two loops of three spells each.

Also, comparing individual offensive spells against individual defensive spells and ignoring mana strengths, it seems that each spell counters two spells and is countered by two spells.

These counters can be represented geometrically if you arrange the spells in a hexagonal pattern based on their mana affinities and the single loop formed by the offensive spells. Each offensive spell does better against the two nearest defensive abilities and worse against the two furthest. See this imgur link. Edit: this relationship is explained by GuySrinivasan who says "It looks like defenses of an element do not protect well against offenses of that element, and defenses of an opposed element protect very well."

Note that opposite mana types are opposite on the hexagonal pattern, so the sine wave component of the mana strength changes can be imagined as a rotation of the hexagon.

Now to apply this to the current situation. Unfortunately, the opponent's offensive spells, Flambeau's Flying Fireball and Merinita's Mud Missiles are opposite to one another. So, any defensive spell that counters one will be countered by the other. 

Unlike GuySrinivasan, I am not eager to experiment with 3 offensive - that's a great idea to test once I'm an official mage, but two risky now imo.

So, I looked at the cases where one duelist picked these 2 offensive spells, and found the win rates of each defensive skill against it, both solo (and specifically in the case where there was only one defensive pick by the opponent) and in each combination.

What I found was that, in the case of picking a single defensive skill against the Flambeau's Flying Fireball and Merinita's Mud Missiles combination, Vaporous Vambrace of Verditius was only a little bit behind the top picks (Citadel and Levee) and will likely be the best once mana strengths are taken into account.

Looking at pairs of defensive spells, the best pairs did better, but not enormously so, and this was with whatever offensive spells happened to be picked. Also, the best pairs tend to include weak mana affinities. I expect the freedom to pick two offensive spells to counter the opponent, plus the strong mana affinities of Vambrace, will outweigh the slight winrate advantage of the best pairs of defensive spells.

So, the next step is to similarly check for the best offensive spell pair to counter Solomon's Solar Shield (and then to try to figure out how mana strengths affect all this....). However, I am going to make a preliminary guess before checking this. The two direct counters for SSS are Flambeau's Flying Fireball and Hylenion's Hammer of the Heavens. Fireball is doing okay mana-strength wise and is an obvious choice. Hammer on the other hand has weak light mana affinity. So the question is, should we sacrifice the counter for better mana type? Obvious choices are Rainbow Rays of Rahl (one strong mana and one weak, not countered by SSS) and Merinita's Mud missiles (one strong, one normal, but countered by SSS). However, this latter choice would make the matchup a mirror except for the defensive skill choice. And I already looked at FFF+MMM's win rates against all defensive abilities including SSS! It did worse than VVV, but not by a massive amount. So, our advantage in that case would be the slightly better defensive counter, plus whatever we gain from defensive mana affinity to stronger mana types. It didn't seem that defensive abilities scaled all that well from mana affinity, and the current strength variations are relatively mild, so I am not sure that the advantage would be all that huge.

So while Flying Fireball+Mud Missiles+Vaporous Vambrace would be a safe bet for at least some advantage over the opponent, my preliminary guess will be to assume that abstractapplic is wrong about mana being more important and to go for double counters to SSS, by picking Fireball+Hammer+Vambrace. Likely to change once I do more analysis!

I also note that the intermediate choice of Rainbow Rays of Rahl mentioned above, if picked instead, would cause my answer to match apstractapplic's "new best guess". We'll see if I end up agreeing!

Edit: I've checked the offensive pairs against Solar Shield and did end up agreeing with abstractapplic. F+R is the second best pair after F+H against Solar and the difference is likely small enough to be compensated for by R's higher mana strength. Still need to study multiple mana strength interactions and mana strength-counter interactions.

My assumption that being able to optimize two offensive skills against the opponent's defense would bring an advantage seems likely wrong (just one of F or H gets good winrate against O as a solo defense if picked with 2 defensive skills), so I may further look into 2 defense+1 offense.

Flambeau's Flying Fireball + Rainbow Rays of Rahl + Vaporous Vambrace of Verditius, same as abstractapplic.

Matching currently-strong mana types is much more important than countering your opponent's choices.

Fireball, Rays, Vambrace

give my master roughly 2:1 odds in favor.

I deduced the existence of Darkness Mana, determined that it almost certainly has a value in the 16-18 range, and then . . . couldn't figure out any clever way to use that information when strategizing. I suspect I'm missing something here.

Wondering about:

Are mana levels actually normalized to 150 total, or is that an effect of our measurement/prediction process?

How do duels actually work? They could be as simple as "add up mana values of elements on offense, modify defense mana values according to effectiveness, subtract; compare results; higher wins" or much more complicated like bucketizing mana levels, choosing a random person to attack first, choosing a random attack and a random defense, doing [something like the simple one] to determine damage, switch attackers, repeat until someone has taken X damage, also wizards have native attack/defense/elemental bonuses, also wizards have different HP so X1 and X2 not just X".

• Based on how stuck-in-the-past all these wizards are,  and how stable the Laws of Magic seem to be, you don't expect that wizard strategies/spell selection/how a duel works/what each spell does should change over time - the only differences you expect between duels are what spells each wizard brought and what the mana levels were.

To clarify, this part of the post is intended as an out-of-game statement to players.  Mana levels are different between duels, but nothing other than mana levels should change between duels - the rules of the duels remain the same.  You don't need to worry about 'what if the rules of duels changed over time', 'what if a better version of this spell was developed halfway through the dataset', etc.