Sounds interesting. Make sure you balance usefulness with difficulty in obtaining a resource. I (personally) dislike it when 90% of items/units are made of (for example) wood, which is really easy to gather, making the remaining resources redundant. Even if those other 10% of items/units are over-powered, players will tend toward mass-producing the easy/weak items/units simply because they're usually able to produce a lot of them before any reasonable number of the 10% would have been able to be produced.

Gold is important for trade and improved resources gathering. Source = Gold Mines.

From experience, tier 2 of any strategy game that allows you to increase your resource gathering is a must. I'm affraid that, if Gold should be usable as it stands, it would become the logical step for most/any competitive players of your game, which would unfortunately tune down the efforts you're putting in these other alternatives.
The second issue here is sciences. Most games with a tech tree make it fairly obvious that this is more than a viable path to success: it is necessary.
Overall, I applaud your idea of letting the player their own strategies, but as the gain from gold and mana is exponential by definition, I'm affraid you only have two realistic approaches from this base economy.

On the other hand, I love that you've catered to the turtle and aggro crowd, this could take a lot of balancing to make it right, but it definitely makes sense that you could have strong fortified positions (stone) vs mobile wrecking armies (iron). Just make sure you do favor defense at all times, as it will balance itself out once you account for mobility.

Also, might I suggest you have a look at Warrior Kings for their implementation of gold?

In my opinion, for RTS games, you would preferably want the games to be as intense as possible instead of a scenario where one player attacks while the other defends. Even worse, the opponent is a turtle(One who only defends) and prolong the game into a stale and boring game. Although turtling is a stupid tactic which can cause you to lose because your resources get exhausted, people still do it which irritates me whenever I play RTS games. Hence, there needs to be a method which can simplify the task of destroying even the strongest defense line. At the same time, not comprimising the viability of light defensive build where players choose to forgo the early game and prepare themselves for the mid/late game.

Method can include long seige weapons such as rockets and tanks which can constantly be a threat for fortification as well as destructive weapons which can only be used once in a while. English bad sorry.

Mine for 100% resources, you could pay 10% resources (gold included) to upgrade an existing Stone Mine's efficiency by 10%. The benefit would be the same, but microed differently

>Strategy games are generally based around territorial control. In this case, I would make upgrading an existing mine less efficient. It is an option if you're stuck on a smaller base, or aren't able to really take the center for example, but sticking to this should put you behind. The reason for this is that you want players to stretch out to locations in hope ot achieving an advantage. This implies risk and decision for the player and all boils down to strategical thinking (estimating opponent forces, reactivity, etc). Works in real time or turn based regardless.

The benefit of the upgrade would be less mines to have to manage, at the cost of vulnerability (losing that one Stone Mine could cripple you).

The reason why this isn't enough is because it enables you to concentrate your forces more easily. If you have 1 mine heavily upgraded and your opponent has 5, its easier for you to pick a small force and harrass each of those than they can (they need to take pretty much all they have and come at you). If the yield of an upgrade has a diminished value (either it takes longer than building an actual mine for the same resource increase, or it is a fraction of the price for a smaller fraction increase, or it has a cap of upgrades, or all of the above) it forces the player to make decisions that factor game mechanics overall.

Science/Magic: Similarly to gold, mana is used to improve your choices. Some bonuses are passive, while others are active (which is why "research" might be a bad name for it - since you'll be dealing with spells, enchantments and potions, albeit functionally similar).

Active is better here, especially if its not more powerful, but mainly for versatility. For example, allowing a pikeman (typically good vs cavalry) to have a shield at all times, which is good vs archers, it just too good. Enabling a pikeman to temporarily lower his guard (lose his bonus vs cavalry and some defense) to move faster (get closer to archers) through the use of some sort of rush/charge command (stamina based let's assume) isn't powerful per se, but if you have a lot of pikeman, and the opponent has only archers at a safe distance, it allows you to take calculated losses to charge into them disallowing them any hope of escaping. Active = option. And they become only efficient if the player knows what he's doing.

In Starcraft, they do manage to pull off a decent tech system, though, because a great micro player can steamroll someone who's going for tech and the upgrades don't give too drastic advantages.

I disagree. By midgame, if you are still on level 0 armor and/or weapons, you cannot compete in any decent diamond or above match unless this is a zerging fest of some kind. Most of the skilled layers end up with victories by outmatching the dmg/def tech. The "other" techs, such as say, Blink for the stalkers, are probably what you should be shooting for. They are options for the most part that you need to invest it, only if you plan on using a strategy that is compliant with that. BlinkStalkers are good, but if you're not chasing air, or don't intend on capitalizing on their mobility to harrass or whatnot, this is a wasted investment.

I'll definitely check out Warrior Kings, thanks for the tip.

It is not a very good game mind you, but it has some good decisions ;)

Even worse, the opponent is a turtle(One who only defends) and prolong the game into a stale and boring game

Most realistic medieval rts generally fail to capture the real essence of a siege. Most pro RTS players do understand the importance of strategy over busting the front door.
Against a turtling opponent, the idea is to cut the chain of supply (Sun Tzu) which means cutting ressources and reinforcements. A player on "one base" can be cut from the outside world. A novice player will buildup an army and strike the coup the grace and will fail (this is how turtling wins: defeating overeager opponents) but a pro will capitalize on his mobility advantage to occupy more resources and control the map.
Medieval sieges weren't so much fighting as movies would like to depict as cutting the castle from supplies and wait for them to starve. Just because you don't get to see the bloodbath doesn't mean its not a strategic win.

Although turtling is a stupid tactic which can cause you to lose because your resources get exhausted

I like a game when its possible to win as a turtle but only if you've planned accordingly. One good examples would be english longbowmen technology employed in the middle ages. It allowed them to strike the besiegers at a safe distance, opening up the supply lines. Their castles would stand unhindered which has been instrumental in many wars.

Hence, there needs to be a method which can simplify the task of destroying even the strongest defense line

I strongly disagree. I feel the trebuchet-like units are hosers for the turtling strategy which should not exist. In reality, the trebuchet fired twice an hour on average (unlike most depictions) and rarely hit critical structures aside from walls. Modern RTS assume that this is a precise tool used to detroy critical positions but it really isn't. If anything, I would recommend to make the turtling strategy more appealing and giving the aggro strategy means to achieve victory through other means:
This falls back to my earlier comments: by forcing players to go out and get resources you form the idea of a supply line and cutting that can be instrumental to victory. Turtling works so long as you can still defend your supply lines.

Method can include long seige weapons such as rockets and tanks which can constantly be a threat for fortification as well as destructive weapons which can only be used once in a while. English bad sorry.Posted Image

I had assumed the project at hand was more medieval-era/fantasy based on the resources chosen, but since you extend me the courtesy of discussing modern warfare:
The dynamic of war makes it hard to do as you would like here, not just because of historical/reality constraints, but because of the underlying logic that applies to these units.
It's easy to think you can rid yourself of an opponent with an A-Bomb, but there's an easy reason why none has been fired in the last few decades: Mutually Assured Destruction (MAD). A-Bombs are not so much a tool of mass destruction as they are an insurance policy saying "If you fire yours, I'll fire mine, and your win-lose becomes a lose-lose, so you have no real reason to fire that bomb". You could argue not all nations have A-Bombs, and you would be right, but nearly all nations have an ally someplay that has an A-Bomb, or even if not allied, there are nations with A-Bombs who don't look kindly on other nations using their A-Bomb.
As you can understand, the political scenery is an important factor here.
Also, while one could remain entrenched in long-ranged sturdy positions (artillery and tanks) the inaccurary of firepower increases the likelihood of firing at civilian collaterals, and that too, is frowned upon the political sceneary. Declaring a war nowadays requires a sturdy casus belli and an ethic of morals/war rules that is much more definite than before.
For these reasons, the "plague" of the 21st century (terrorism) is harder to deal with. You can't A-Bomb terrorists hiding in an otherwise perfectly fine country as the entire political scenery would point the finger at you for the civilian losses and lack of judgment. You can't come with tanks in the middle of a city that you suspect to host a few bombermen. What you can do, is bring in the G.I. which are a lot more accurate, but shorter ranged and less armored.
So modern warfare would be more like this (a strange rock paper scissor)
- Infantry (high losses for the player, short range, high accuracy, rarely kills civilian, politically approved)
- Tanks (low losses for the player, long range, medium accuracy, kills some civilians and economic structures, politically ambiguous)
- Artillery (low losses for the player, longer range, low accuracy, kills civilians and economic structures, politically shunned)
- Missiles/Bombs (N/A, long range, accurate, deadly to civilian populations, politically loathed)
Based on your actions against one nation, the political scenerary adjusts, and the next time you declare a war against an opponent, based on your previous actions, other nations accept or not your casus belli (based on its strength) versus your acts of war.

This goes for a very complex game, but is closer to depicting the actual downsides of using long-ranged units, because let's face it, if there weren't downsides in real life, we'd all be A-Bombing one another and infantry just wouldn't get used.