The past dozen posts are great.  I never wanted to implement an exact copy of the SMAC technology system and I apologize coming across that way. I was trying to expose the idea of a non-standard tech system to the rest of the forum and SMAC was the best example i could think of. I was also trying to stay away from specifics b/c we are operating in near total ignorance of what the developers are thinking about technology at the moment.

I think a situational bonus to research is a great idea but in many cases it was not an abundance of resources that lead to advances but the scarcity of them. For instance, the technology for the steam engine existed in ancient Rome but there where so many slaves that a labor saving device simply wasn’t needed. Another, is that if you have only a small amount of a metal you will only allow it to be worked by your most talented people thus possibly producing higher quality goods but then again practice makes perfect.

Overall both systems are very intriguing and I am very happy about the direction the conversation has taken. I need to the past posts a bit more carefully before I respond more than this. Great ideas

Excellent posts, some food for thought. I really loved the SMAC style of research. I'd have to come down on the side of 'blind/focused/specific' model. Offers great way around the 'blind' V 'min/maxing'. How about including an ability or skill that your channeller can choose at development (gain through levelling or build specific building) that gives an ability to 'see' several choices ahead on a research tree and adds a bonus to researching that path (i.e. costs mana to 'see' ahead)? Or too much feature creep?

Great. So it looks like we have a consensus. Now for some feedback from SD...

Personally I don't think that the tech tree should be hidden from the player. A hidden tech tree would drive me crazy. It could be fun and immersive the first couple games, where you're really kept in the dark about where your research will ultimately take you, but that novelty would quickly fade for me, I think. Also, people would just map out the tech tree and print it out, making a hidden tech tree nothing more than a burden. I'm definitely all for a visible tech tree, regardless of what style of research is implemented.

On the other hand, maybe there could be a Divination spell that would function differently depending on how you use it. Using it on another town or army would amount to spying, using it on your own lands could reveal hidden effects/resources/dungeons or something, and using it on your research could give you insight, allowing you to choose a specific technology to research. Using this would require essence, so doing it every turn would result in a weaker channeler so that if you really want to min/max your techs you can - but at the price of a weakened channeler. On the other hand if the focus system is done well, this might just contribute to feature creep without improving the actual game.

But yay! I created an idea that people like!

I'm glad it looks like we've a palatable solution.

Ok, let me restate the goals here:

- make it fun for those who want blind research
- make it fun for those who want to pick specific techs to research (myself included)
- don't place undue or unnecessary burden on development/testing/balancing resources (e.g. avoid options)

And a proposed solution:

- when choosing what to research, the player can choose between three methods:
-- Pick a specific tech like normal (and yes they can see the whole tree); this will be, maybe, 75% efficient
-- Pick a tech category, and the computer randomly selects an eligible tech in the category; this will be, maybe, 90% efficient
-- "Blind Research", and the computer randomly selects an eligible tech; this will be 100% efficient

That's it.  We can add complexity with character bonuses, situational bonuses, etc... but this is the kernel of it I think and I'd rather stick with simple until we actually get to playtesting.

This seems like it would add fun value in that it:
- poses a tradeoff; no one of the three method will always be the right one.  If you do 100% blind you will have more tech points worth of techs, but it probably won't fit any particular strategy you could execute.  If you do 100% specific you will probably fall behind in tech overall even though you have bee-lined to what you want.
- since accepting a degree of blindness actually helps maximize research output, min-maxers like myself will have a reason to build strategies not based on a lock-step progression of techs

The three efficiency values (75, 90, and 100 or whatever) should be stored in the main "global settings" xml file or somewhere similar for easy modding.  That way other values (like 100,100,100 or 0,0,100) are available to the player who just really wants it that way, but testing/balancing resources aren't used for the alternative setups.

Sometime later I'll post some nerdery about what I imagine it will actually take in nuts and bolts to make this happen.

Thanks for the discussion everyone,
Keith

Excellent summary, Keith. Thanks for the doing that.

I would also concur that wrinkles like the situational bonuses can be left out until we see if/how the core idea works after some playtesting and balancing.

- Ash

Ok, as far as what it would take:

1) add three entries to the settings xml file:

<Tech_ResearchEfficiency_Blind value="100" />
<Tech_ResearchEfficiency_Category value="90" />
<Tech_ResearchEfficiency_ExactTech value="75" />

2) add a new data type for "tech category", the only fields I think it would need are an ID and a name, but perhaps the image/graphical-resource would need to be on there for the button that will be clicked to select it

3) add a "category ID" field to the "tech" data type that will link each tech to a single category

4) add two fields to the game-state data type of a player:
- "Tech_ResearchingBlind" that stores true or false
- "Tech_ResearchingCategoryID" that stores a categoryID (blank or null means not doing category research)

5) on the choose-tech screen, add a button for "blind research", and a button for each tech category (or to make it mod-friendly, just a drop down select box with the names of each category)

6) when blind research is chosen, pull the list of eligible techs and set player's tech-in-research to one at random, and set that player's Tech_ResearchingBlind to true (and Tech_ResearchingCategoryID to null)

7) when a category is chosen, pull the list of eligible techs in that category and set player's tech-in-research to one at random, and set that player's Tech_ResearchingCategoryID to the selected category (and Tech_ResearchingBlind to false)

8) whenever displaying "current tech research":
- if Tech_ResearchingBlind is true display "Blind Research"
- if Tech_ResearchingCategoryID is not null display the name of the category
- if either Tech_ResearchingBlind or Tech_ResearchingCategoryID is not null, display a term based on percentage complete divided by 20:
-- 00-20% done: progress bar displays "Imagining".
-- 21-40% done: progress bar displays "Hypothesizing".
-- 41-60% done: progress bar displays "Testing"
-- 61-80% done: progress bar displays "Theorizing"
-- 81-100% done: progress bar displays "Verifying"

9) when processing a turn and figuring out how many tech points to add to a player's tech-in-research progress:
- compute research points normally
- if Tech_ResearchingBlind is true, then multiply result by Tech_ResearchEfficiency_Blind
- if Tech_ResearchingCategoryID is not null, then multiply result by Tech_ResearchEfficiency_Category
- if Tech_ResearchingBlind is false and Tech_ResearchingCategoryID is null, then multiply result by Tech_ResearchEfficiency_ExactTech

Am I missing something?

(and yes, I realize this is probably overkill, but it helps to work through something rigorously to check for surprises and maybe even save the devs some time)

Thanks,
Keith

I concur on both points. Moreover the summary is worthy of karma.

I was going to suggest that. Great job! A classic case of out-of-box thinking.

One small point I would consider renaming "category" to 'general' and "exact" to 'specific' for consistency purposes.

Yea, I was thinking that magic research should stay MoM-like in that you pick a specific spell out of a spellbook which already has the thing it just needs to be decoded and understood.

Awesome thread.  Figured I'd add my thoughts, coming from a guy who has never played SMAC or MoM.  My first preference would be to have an option at start up.  I guess I just don't like the thought that if I specifically pick my techs, why should I be penalized and potentially way behind someone else in techs in the late game?  What if I just like to pick my techs, without doing the whole min/max strategy?  Does that mean I'm also getting penalized although I'm not really getting ahead in any one thing?  I'm kind of a jack of all trades, master of none type of guy so it seems a little unfair.

I would like a system where specific research is only penalized if you keep on down the same path.  For simplicity's sake, take GalCiv2:  If I research laser I & II, there is no penalty, but if I go straight to laser III or above, there becomes a progressively higher penalty if I'm not researching other techs.  Obviously, it gets more complicated with prereqs and a less linear research structure, but hopefully get my general idea.

That said, I think the category research is one of the better ideas as I'm too used to specific research (think GalCiv & Civ4) and it would add to the variety and replay value of the game.   It'll be interesting to see what Stardock already has in mind for research.

I can happily agree to the following;

Goals:

- make it fun for those who want blind research
- make it fun for those who want to pick specific techs to research
- don't place undue or unnecessary burden on development/testing/balancing resources

Proposed solution Core:

- when choosing what to research, the player can choose between three methods:
-- Pick a specific tech like normal (and yes they can see the whole tree); this will be, maybe, 60% efficient
-- Pick a tech category, and the computer randomly selects an eligible tech in the category; this will be, maybe, 80% efficient
-- "Blind Research", and the computer randomly selects an eligible tech; this will be 100% efficient

The three efficiency values (60, 80, and 100 or whatever) should be determined through playtesting and stored somewhere for easy modding. 

Additional Modifiers outside of current scope: character bonuses, situational bonuses, etc...

Sammual

I didn't address the AI part of the implementation of the blind/category/specific idea.  I think this might be a sticking point.  At least, it would be the most complicated part.

A few solutions come to mind:

1) just let the AI do specific-tech research at 100% and not be aware of the blind or category options

- con: the AI would not be playing the same game as the humans, which is generally a big no-no particularly in Stardock's history.

- pro: dead simple solution, and it probably wouldn't give the AI much of an advantage since AI min/maxing pales in comparison to that of some human players

2) build logic that takes the normal pick-most-important-tech functions and expands them to work with the new system:

- calculate an "how-much-I-want-this" value for each tech that can be researched now (this is the hard part, but they need thisfunction for specific tech research anyway)

- let blind_value equal the average of how-much-I-want-this for all eligible techs, multiplied by 1.33 (if using 75/90/100)

- for each tech category

-- compute the average of how-much-I-want-this for all eligible techs in this category

- let category_X be the category with the highest average how-much-I-want-this

- let category_X_value be that category's average how-much-I-want-this, multiplied by 1.11

- let specific_tech_Y be the individual tech with the highest how-much-I-want-this

- let specific_tech_Y_value be that tech's how-much-I-want-this (no multiplier)

- if specific_tech_Y_value is greater than category_X_value, next research = specific_tech_Y

- else if category_X_value is greater than blind_value, next research = category_X

- else next research = blind