There are several problems with that theory. Firstly, by your own admission, it requires a high degree of fortune for optimal performance. If you've played strategy games in competitive settings before, you know that the way to win is by employing methods that produce stable and relatively predictable results... essentially the difference between playing a game of Texas Hold 'Em vs. playing cee-low dice in the alley. The key to success in strategic gaming hinges on your ability to accurately assess your strengths, your opponents strengths, and the general outcome of the various options available to you, and based on this knowledge, choose the best path. This is just plain impossible with a strategy that relies on randomly generated quests and rewards.
Secondly, you completely mischaracterize the strengths of someone pursuing a Warfare/Conquest research focus. They will not have "peasants armed with clubs" by the time you have reached LEVEL 4 spells at the minimum to get decent summons! Even if you rushed straight through the levels without pausing to research a single lower level spell, you will be outpaced, and your opponent will be generating 8 figure troops with warhammers. How you would even survive long enough to summon your stone giant (who, by the way, can only fire his ability once every three game turns) is beyond me.
Third, and lastly, where are you finding enough elementium to reliably supply even one multi-figure unit, let alone attempting to keep pace with someone that can crank out troops restricted only by metal. What game are you playing that heroes are unstoppable? I suspect the answer is "Elemental, single player, with autosave every turn and a reload every time one dies." Heroes are weak to the point of uselessness, unless you are using the bugfix mod which enables shard spell damage multiplication as it was supposed to work at release. Then heroes can crank out good damage if you find a shard aligned with your spheres and imbue them.
That's just a gross misrepresentation of facts and absolutely impractical given an opponent that is playing to win.
Perhaps, instead of arbitrarily limiting narrowly focused research, it just becomes more difficult by a certain % for every advancement made in a row, starting at 3. I dunno. But, and soon we shall be able to test, any player in a serious game that does not rush warfare will get curbstomped before turn 75.