I've never really been sold on the idea of dragons - they can be useful, but I honestly don't remember the last time I saw a dragon last longer than 3-4 hours. I've also found it a little silly that the enemy kingdom already knows what you're going to send as soon as the project begins - a well organized kingdom can kill it in an hour if they decide they should, or they can even choose to ignore it.

I would like to propose a revamp of the entire dragon system, including ways to "heal" a dragon.

Customized Dragons

1. All dragons will have a base cost, using the same formula as they do now.
2. All dragons will maintain the base damage to income and draftees.

3. Final price will actually depend upon the options chosen:
a. Fear: increased military losses.
b. Timidity: decreased land gains.
c. Paranoia: building effectiveness reduction.
d. Armored Terror: decreased OME/DME. (possibly separate options)
e. Hunter: decreased WPA/TPA. (again, possibly separate options)
f. Production Pauses: decreased food/rune production.
g. No Peace: decreased birth rates. (10-15%)
h. Destroyer: chance to destroy buildings per tick. (50-75% as powerful as a single Tornado)
i. Strafing Run: chance to kill peasants per tick. (50-75% as powerful as single fireball)
j. Horrifying Presence: increase wages by up to 15%
k. Strengthened Hide: increased overall life. (25-50%)
l. Regeneration: chance to regain x% points of life per tick. (25-50% chance, 5-10% life regain)

4. Either a slider or point system could be used to determine the severity of the effect and cost of the effect.
a. Yes, you would be able to max out ALL of the penalties, but your dragon will probably cost more gc than an entire kingdom can generate in 72 hours. But it could be done.
b. Yes, you could add the minimum of each penalty.The dragon would be cheaper, but not very effective.

5. The target kingdom would also not know anything other than "a dragon is coming" until it was actually sent, and then the dragon page would reveal the powers and strength of the powers.
a. "Kingdom X:X has begun a dragon project against us! We will not know the strengths of the beast until it arrives."

6. There would obviously need to be caps on the penalties the dragon could apply.
7. A slider type system would allow for fine-tuning of strength vs. cost.
8. A point system would create an artificial cap, based on the cost of the penalties to impose and the default points available.
a. Perhaps more points could be purchased for an increased base cost.

Example: (Rough example, I know this isn't the dragon cost formula.)

Drake: 35 points, costs 90% of kingdom NW
Adult Dragon: 50 points, costs 105% of kingdom NW
Ancient Dragon: 75 points, costs 135% of kingdom NW

Fear: 2% increased military losses - 1 point
4% increased military losses - 3 points
..
..
10% increased military losses - 10 points

And so on.

The numbers are pulled from thin air, they are in no way representative of anything balanced. Very rough idea that probably needs a lot of refinement.

Also, dragons can be named.