I like the idea of research effort (for spells/tech/whatever) being somewhat uncertain, especially as the amount of user made content grows and you might not even know what all of the potential spells are/do, but I'm wary of it as well. I think my concern would be that you might get shafted with no way to recover (through no fault of your own), which could be especially annoying late game.

What about an adjustment of the idea - you're researching "the next spell tier" in Water+Life. You don't know exactly what spell you'll get, but you know you'll get some spell from a group that's equivalent in power. So maybe you get Nature's Pity or maybe you get Nature's Wrath (or some other power-equivalent spell).

Or perhaps your research reveals that you can't get the spell you were trying for (Nature's Pity), but through all your effort all of your Water+Life spells are cheaper to caste since you now know so much about the topic.

I'm just not sold on the idea that I could be researching a tech forever (in a game, mind you) and never get anything for the effort.

And yes, I often turn off the Mega Events in GalCivII for that very "late game getting shafted sucks" reason Sometimes it's fun, but more often I feel frustrated after investing hours of time that I'm now in a signficantly (and maybe unwinnable) scenario.

I'm not a fan of complete blind research.  Its not intuitive and its not all that natural, and generally makes for disappointing sessions when you get a string of bad luck with your dice.  Yes some discoveries have been made on accident, but its hardly the rule of scientific advancement.  As Tamren said:

What could be interesting would be 2 types of research, experimental and advanced research.  Advanced research, is simply the advancement and improvement of already existing knowledge, basicly if you have fireball lvl 1, you can research fire ball lvl 2.  Experimental research would then be discovering new spells altogether simply by trying new techniques and ideas.

Experimental research wouldnt be completely random though.  If you had your fireball spell for instance, it could serve as a knowledge base for a firewall spell. But if you had no knowledge of a certain area of magic, it would be mostly blind when you research your first spell in that area.  Even then, spells could be cross connected even amoungst different elements.  For instance, a firewall spell could lead naturally into a wind wall spell.

Its all a matter of how elaborate the connections are.  Now, I'm assuming EWoM will have your standard, click and research method most games have, where you pick your way across the tech tree.  And will have an option at game creation for blind research. I'm not saying they wouldn't want to make a more elaborate system, but chances are, the budget and time constraints will require them to focus their resources else where.

I really like DrGuppies idea, you indicate in which direction you research but you can't say which specific spell will be found.

I definetly think that you should be able to direct your research....

How much does it suck to be trying to research an offensive spell (I am researching fire here....) and then you end up with something that you didn't really want--- Wall of fire... 

Its nice and all.. but now you have to start out again trying to research.  That always feels like a let down.

I do like the random chance of success though.  Both MOO2 and MOM had this element.  each turn you would get a certain amount of guaranteed research points as well as some random amount.  Either that or you have a certain amount you gain each turn and then you have a chance to succeed... 1.5x total required = guarantee, less and you get to roll til you get there.  That way sometimes you get lucky and get a spell early.. sometimes you might have to wait a little longer than normal.

Of course, the random element shouldn't be too big. But some variety in the spell research tree (or at least in the order you can research them) would work fine. Cfr. MoM where you also had a random factor determining which new spell became available next.

There could be some more intelligent mechanics rather than just a random factor, e.g. if you could research "offensive fire spells" you don't know exactly which one will come out but at least you determine the direction you want to go in.

Sounds like the same as AOW:SM... as I mentioned earlier.

some randomness in games makes the games more fun to watch and play.

But it comes at the price that the game can never include a realistic ladder, since a relatively poor player can quite often beat a top player. Aka master of magic, AOW series etc

I also thought of a second way for research to be done, the could satisfy both the desire for randomness, while still keeping victory to skill and not blind lunk.

The basic structure is a bit simplified over all.  Instead of researching a specific spell, you research a school, and advance your knowledge in that school.  There are a few basic tried and true spells, that you will always gain access to in each school at every teir of advancement.  At the same time, you are given a random selection of more complex spells that are more situational in nature of which to pick from.

The idea, is that the spells that are more generally useful and a bit more boring are always available.  And the more situational, varied and interesting spells randomly become available if you're lucky.  You might get a bit of fortune and be able to choose a powerful and fun spell if not limited, but reguardless you'll have that fireball spell, or general healing spell at your disposal.

Luck always plays a factor - even if it is that you make a lucky guess as to what your opponent is up to. Luck should never be so dominant that a poor player beats a skilled player.

Yeah its hard to decide which would work better because we so little about what is planned.