I think that is where we have had a communication error.  Yep auto generating tech trees is a much better way to put it.

I never wanted the core techs to be hidden.  They should be visable.  The upgrade techs, and the fancy extras should be hidden.

Core techs (fixed tree placement, visable.) - bowmen, calvery, farming, etc.

Advanced techs (variable tree placement, hidden until unlocked.) - crossbowmen, heavy calvary, irrigation.

Excentric techs (random placement, hidden until unlocked.) - tech bonuses to core / advanced techs, healing herbs, etc.

as far as realism,  the list of selectable techs would just be ideas that your people have had, and you chose which ideas you want to invest in.

I'm feeling realistic.  Multi-faceted technological breakthroughs, you research woodworking and improve the cost of your pole arms, the strength of your bows, etcetera.  You guess at what it actually did and go hunting for changes if you want to find out.

Randomization is annoying me more and more as I hear the crazy ideas some of you have for it...

What I find disappointing about the techs presented in the july faq is that apparently every single race - or whatever word is more applicable in EWOM - can use every single mount as long as the right tech is researched. Imo it would be way more fun to have each race linked to different mounts - maybe in the form of a bonus like 'race A gets a +1 combat bonus if riding wolves'.

What I do like about the teching is that it potentially opens up different options to winning this game without the need to actually having to plow through hordes of enemies. One of the things I disliked about Civ IV was that it was such a huge hassle to mob up the AI that were left when the game was in a winning position. MoM had a better sstem for this because you could attack just the wizwards fortress and get an opponent out of the way and I hope to see something similar in EWOM.

Apart from that I see different people making different statements about how they see the tech tree working. Most of them sound fun and workabe, yet I do not favor any one in particular. Once Stardock proposes different systems that we can choose from it may be more viable to discuss what works and what does not.

I really wish research was more like random discoveries instead of research a specific thing. How many leaders go into a lab and say ok I want you to research lasers with X components and Y disfusions and Z hyper active expulsions?

Research should just be a population value, education of that population value and costs of resources and of course random intelligence of those educated population values and then let the scientists or in this case the research wizards just have their go at whatever comes.

Many discoveries in our own world were by ACCIDENT when trying to actually research something else. Like the microwave I think was an accident and washing detergent and taffy.

I just wish there were more "surprises" in research instead of the cold hard facts in 20 turns you are going to get this upgrade. That is sooooo ummmm bland an unrealsitc really.

Master of Orion did this well I think where you put in so many research points into several catagories or just one catagory and you got potluck on what you discovered over time.

Worlds should be unique each game not the same each game. Thus there needs to be more whatifs and surprises and so what if you got screwed this game and didn't discover the sword of everslaying the first week of the game. My gawd you can play 10000 games of it afterwards. lol SO what if the AI borks you one game because it got lucky and you didn't? Must every game be a I WIN or the game suks? Must every game be Balanced in your favor or the game suks? You don't win anything when you beat these games anyway so why not opt for more challenge and harder ways to win it instead of WAH the AI got the Sword of Everslaying before me and that's no fair WAHAA!

Well I'm all for anything but canned and knowing what you are gonna get research. Forrest Gump said it best; Reseach is like a box of chocolates you never know what you're going to get.

Can't you solve the scale problem by basing the bonus off the DIFFERENCE in techs?  Instead of 'sword 27 = very slightly better than sword 26', 'sword 27 = effective vs armour 27, less effective against armour 28 and very effective against armour 26'.  In this way, techs are like 'skills', and it doesn't matter how good your armour is - only how much BETTER it is than their weapons, for a floating bonus different for everyone's tech level without needing a system whereby it's nearly a waste of time doing further research.

Oh yeah, the costs should still increase in a linear way, so that to catch up to 'swords 27' is non-trivial.  But rather than having endlessly diminishing returns of +0.0002 bonus, just use the difference as a modifier.  This also would mean that getting swords 27 when everyone has armour 20 probably isn't worth it, as the bonus/cost would be bad at this point.  Once everyone else reaches swords 24, it's more worth it.

No, because your power would be an S-curve instead of an exponential curve, with the inflection point where weapon tech = armor tech.  Interesting idea!  It prevents high-tech from ever being overwhelmingly powerful, but still makes it important to stay ahead in the tech race.

However, it would also be a little difficult to carry out, requiring all kinds of tables and relationships between different techs.  Easier to screw up, I imagine, and a little risky simply because it's novel (I think).  And furthermore, it leaves open the problem of a tech 27 weapon versus an enemy (like a wild grizzly bear) that does not wear armor.

It might be that this model better describes reality compared to the standard model, since weapon and armors evolved competitively and neither really became exponentially more powerful.  Instead, a weapon might arise that is no more powerful against ordinary flesh, and no better against armor, but is better able to disable shields (like Roman javelins, whose name I forgot).

For the infinite-tech weapon/armor stuff, you could take the HoMM route to Att/Def. Every attack over the enemy's defense is +5% damage, every defense over the enemies attack is +5% Extended HP. Basically doubling damage with +20 attack, and half damage with +20 defense. As a second example, 32 attack rating vs 30 defense rating does +10% damage. I don't see why something similar couldn't work for Armor/Weapon tech lines, though it is a little dullish for a research.

As for random tech trees - I find these perfectly fine if done right. I play Sword of the Stars quite a bit, and the most important things are baseline (subjective opinion, some of my friends disagree), everything else is a 1-100 dice roll with the odds pending race. It's extremely rare that I've felt failing a roll has had a significant outcome on the game - though I suppose it can be if you openly rely on something that has a random chance to begin with. More often than not I simply find some other way to press an advantage.

Taking the SotS example a bit further: Techs were NOT cheap, the money your enemy spends on a tech can be money you spend on combat units (often a LOT of combat units). You can then use the ships you have to cripple their infrastructure before they can even field their new ships. The Zuul in SotS were basically built on this, having both poor research efficiency and a high failure rate on acquiring techs to research in the first place. In other ways the the advantage of tech was slightly diminished by high-tech ships being significantly more expensive to build.

I wouldn't want to copy and paste the SotS tech tree style by any means, but I think it's still a good example of a random tech tree that can be learned from. Ultimately I feel a properly done random tech tree can give good players a way to show solid planning and creativity over playing spreadsheet general and making a static order of things before even starting the game, while still at least providing enough strong core techs that can't be failed as fallbacks to ensure a few bad bits of luck don't cause everything to go bad. Personally, I would far prefer that the techs available be random, and not the techs I get upon researching something.

Edit: Now with paragraphs!

I wonder if an absolute differential makes sense with infinite research, vs. a relative differential.  That is, if Attack = 10 and Defense = 5, rather than say +5 we could say 2:1.  Thus, as kingdom/empire tech increases, the absolute difference matters less and less, but always provides at least some edge.  This tends to make a logarithmic curve (which I prefer) but keeps things a bit simpler to evaluate.

Thoughts?

NO! Clearly "Sharpened Pointy Sticks XVII" Is a better upgrade because you get a better bonus earlier. I know you need to research "Blunt Objects" first but trust me pointed sticks are the way to go.

That said I prefer infinite slightly unpredictable research. I'd like to give a command of "better infantry wielded ranged weapons" and get surprised with either a better bow (longbow) or a whole new path such as the crossbow. Then later be able to give a more specific upgrade order such as Improved Crossbows. The idea being without the "pure research" being ordered you rarely get anything new but by ordering pure research you can get a revloutionary technology. At the same time I'd like to see a mix of pure research and specific research.

By the way Saber the fix for a monster that doesn't ware armor is simple. You give world spawned monsters a diffrent class of armor. That class of armor would be designed to scale based on the level of the creature and the weapon being shot at it thus allows said monsters to remain a threat.

Haven't played much of the MOO series, and it's been years since I have and don't remember the randomized tech mechanic, but I'm with many of you in thinking it adds an additional layer of variety to a mechanic which could become stale.  Someone mentioned Civ 4.  In that and many other games one not only ends up memorizing the tech tree eventually, but their strategy tends to follow the same pattern from game to game.  At least mine did.  Randomization helps by keeping the player in that "explore" stage of the game where they're not certain how a strategy will work and you still have to noodle through your decisions based on what’s available.  More like a card game.