@Wizard1200 - I like the general concept, but damage needs to scale per level and not be so tied to shards. Doing 6 points of damage against anything unless you find shards is nasty. It isn't as if we can train the mage in schools after the start of the game.

The way that the game is currently working, it makes sense to tie new spells to research like they currently do, but damage or affects scale according to how you have built the wizard, i.e. water two means you get + 2 levels to what ever 'water' type creature you summoned or +1 to area of affect, with specific spells tied to schools, like they currently do combined with the element school level as you are suggesting to whether or not you have trained high enough to cast it.

Go to Prociponee (if she is in the game) and get the appropriate books. You can train the mage into any school you wish.

Yes, the damage spells should be only one part of the mage and he should able to do many more things with his spells. For example mind control that works like silver tongue should be level 5 water spell or the teleport ability of the thunderstrike spell should be a seperate level 5 air spell without a casting time and without a cooldown.

I think the scaling per level could work if the player selects abilities that improve his spells. For example expert aeromancy (minimum level 4) could increase the number of air shards by 2 and master aeromancy (minimum level 10) could increase the number of air shards by additional 2.

I would really like to work as a freelancer to improve the magic system of legendary heroes by the way

Incidentally, is there any way to learn life or death short of sacrificing a hero that has it?

Not that sacrificing a hero is a particular hardship in this case. 

So, um, release today? PLEASE?!

Thank you!

Why the nerf to Warlock?

It is already pretty mediocre. How many direct damage spells are there? 

Most Sovs start with zero.

Come to think of it, the whole magic system feels a bit clunky: a profession (sum or warlock) and magic schools at creation, followed by path of the mage, with more spells and traits similar to summoner and warlock.

The spellbook remembers the last strategic spell you cast and starts there when you reopen it

Thank you! These are the quality of life issues I like seeing fixed. I can't wait to test out the new build! Any eta yet?

Fantabulous!

I can now upgrade my units squad, oh happiness!

Except for Life - you can't add life magic. 

If we get road building it has to be done by worker/pioneer units this way. Of course they could add an automatic function for them like in Civ5 and give the workers a route to follow. It's alot more fun when you have workers that can be attacked by enemies and monsters. If it's just a build road from here to there order with no worker unit, what's the point in changing the current system afterall?

If roadbuilding is implemented properly, there also need to be an upkeep cost to them, so the whole world isnt spammed with roads. The benefits are faster movement and the trade routes/caravans. I also think that building 2-3 tiles of road per turn is far too quick. 3 turns for one tile on normal speed is more like it.

I think this succinctly describes everything that is wrong about manual road building.  I hate it with a passion.  I am so grateful to not have to deal with the mundane minutiae of roads.

I think most of the issues with road-building would be resolved if the local road system were recalculated every time you found/take a new city.  It would be a little weird to have the roads redraw, but hey, roads are already magical abstractions.

Waiting to see a patch note about this, but the way Heart of Fire applies it's attack bonus to each "figure" instead of per stack is a bit ridiculous.  +15 attack to a 5 man unit from a no upkeep spell is a bit over the top (3 essence city, could be more or less)

Also, the "closer, faster" combat is great, but most combat magic is uselessly weak, overexpensive manawise, and generally boring - particularly the spells that take multiple turns to cast.  Out of my 4 games, I have never been happy that I wasted the mana on the weaksauce that is Fireball, considering my Horsemen of Doom have typically already won the fight before it finishes casting.

What are the issues "issues with road-building" that you mention and would be resolved if the road system was recalculated with new cities?

It is intended that the road system grows out of the way your empire grows.  That it may start with a single road between your first 2 ciites, then expand as additional cities are added and that it prefers those existing roads that are already in place.

We tested having a road from every city to every city.  It was a mess, basically every tile became roads (because the most efficient connection between 2 cities in each case resulted in lots of roads).  It looked horrible and it wasn't fun.  So the current system prefers existing roads so you get more natural looking roads.

Definately agreed. We don't want to have it looking a mess. I think two main problems exist that would eliminate 90% of the concern.

First, when an auto-road is placed, there are times where the auto-placment of the road prefers a route through wildlands or through a dragon's lair or around a lake or something like that even though it's clearly a longer route compared to what the player would tile out.

Second, when you have a scenario when all cities are connected to the road network, but the orientation of that cities warrents another road to be built.

A----B
       |
D----C

Why would no road exist between A and D when D is built?

Understandably, the Engineer unit or Capitar or Commander's have been considered or implemented in order to help address these two conerns. However those are bandaid solutions. It doesn't fix the initial problems. It just gives you the option to eventually work around it. More over, it comes a matter where the player HAS to work around it in some cases which forces their game down a path the player did not want it to go. Likewise, the AI does not understand how to utilize these bandaid solutions in the same manner as the player and as such are placed at an immediate disadvantage.

Painfully, I am no programmer, so I cannot offer advise in how to fix these concerns. All I can do is comment as to what the concerns are.

ps. if a solution is ever found and implemented, PLEASE backport to FE.

Thanks GFireflyE.  Your example is perfect.

If, as you stated the game was to always use the shortest distance, then not only would we get a road between A and D, we would also get roads from A to C and B to D.  We would have a huge sprawl of roads.  We had this, it looked horrible.

So instead what the game does is it costs the movement between A and every other road with the existing roads.  If it is ever shorter to go onto a non-road tile to connect any of those cities, then it buidls a road there.  Since our roads give such a big movement advantage, its pretty rare (honestly if were were to reduce the movement bonus on roads you would see more hookups between cities).

So in your example when you place city D, city D finds the fastest trip to cities A, B and C.  In your example it is faster to go along the roads from D to A then it is to go over land from A to D, so it uses the existing road structure.  Granted if there was a road between A and D then that would be even faster, but that would cause the road spam I talked about.

If A and D were significantly close enough to each other and far enough from everyone else that goinging overland between them would be faster than taking the road from D to C to B to A, then the game would have placed a road form D to A.

In one game I played I had this happen but I didn't control B or C. I started with A and conquered D yet no road was built between the two. Very annoying since it took like 12 turns to walk between the two.Other then that I can deal with the current system.

Derek i think you can use a better algorithm then the one you mentioned for building roads; an algorithm which will prevent spamming of roads and yet provide a coherent network. 

First of all, you make a road to a closest city - we will call it B. Thus, we have connection A-B. Now, we start to check other cities in vicinity of A (vicinity, let's say 2 or 3x distance of A-B  ). Let's say there are cities C and D close to A.

I check distance from A to C through B ( A-B-C) and compare it to a direct distance A-C.

If A-B-C>2*A-C then i build a road A-C. (In other words, if the direct road A-C is twice shorter than long connection A-B-C a road is build. )

Next, step - i look at city D and follow the same pattern: compare A-D and A-B-D and if A-D is twice shorter then i build a road. 

In the previous post square example, when city D is build it will connect only to A and C, and not to B creating a spam of roads. It will also allow that if you build a outpost between two cities a road to both cities will be build. But it will prevent spamming a connection - all cities to all cities. 

Wow, this is a great list of changes. Thanks. Definitely seems like Stardock is taking advantage of forum feedback.

Not sure I understand the place of swords now that they've been nerfed even further (-4 attack to longswords, wow), but you made axes and maces really distinct in LH, so I'll assume I need to tinker around with swords to find their strategy.

The main place for swords is defenders, those unites you put on the front line and set to guard stance so when they attack they counter. With the +defense while defending they are tough and counter to do damage those that attack it. At least that's how I see it.

with the recent change, all swords have at least +2 initiative, so it's probably a good idea to design sword units with the finesse trait to make use of their higher initiative. possibly also the "fast" trait. they don't get the fancy AoE/double damage attacks of the other melee weapons, but swords are now reliably high initiative weapons (unlike the current system where the final upgrade removes the +2 init of its predecessor). not sure i will actually bother with swords, but at least the whole category is consistent now. a unit design with sword, finesse,fast will consistently move first and get the bonus damage vs. all other trained units (unless you give them plate armor and let them fight units with no armor - which is mostly irrelevant since your guys will still kill them without taking any real damage).

^ This

But you can't go too far with the algothym or it turns into connecting all the cities to all the cities, which is what Derek is quite understandably trying to avoid.

In addition, when you conquer a city that already has roads put in place, roads are not built towards your empire. This needs to be implemented, perhaps I misunderstand what's happening. Perhaps it's connecting to the closest city still, but that city is owned by another faction. If that's the case, then the algorythm should should be connecting a newly conquered city to the closest city that you control.

Direct distance isn't a very good indicator because it fails to take things like mountains and lakes into account.  2 cities may be close but separated by a mountain range, and it may make sense to use existing roads between them rather than create a new one.

"Make sense" may be the wrong term.  The hangup is in finding the balance between to few roads and missing some roads in some situations where players want or expect them, or to many roads and having it look bad.

Ahhh the fun of path finding algorithms and the optimal strategy of creating them... As Derek points out it is not a simple or even easy problem to solve.

Also, to lead into more difficult senarios, what if someone uses raise land or lower land and create new paths for the computer to take that into account. Should the old roads be demolished in favor of new roads or new ones created to represent the change in landscape. What about the idea of a volcano going over the current road system?

The random nature of the map leads to road problems. I'm thinking that the road problem solution is not going to be straight forward as people present. I can set up a series of equations to provide the best algorithm for placing roads, but I think that the current system in place seems the most organic. I would be more pleased with the road system, if there was a late teir tech that provides the ability for the roads to build bridges over rivers...