Two assasins

[LPhillips]

As for splitting the skills, simplicity is the goal. The defense is the real problem.

Actually, I think it would make sense for ambush to work on defense, and ONLY on defense.

Applying a bonus balanced for 8 units to as many as 32 is where things go wrong. Ambush is purely a game of numbers: how many rolls vs how many troops. Assuming the opponent employs an ambush hero with a very few ghosts (50% ambush?) in a city, then his real army will still be roughly equivalent to the enemy's, and his static defense will be impeccable without significant cost to his overall strength. Sure, everyone can do it, but it creates a boring, static, frustrating game in which other tactics remain ineffective.

We can't simply make all maps around the Ambush defense strategy in order to mitigate an out-of-control mechanic. However, I think I have a balanced and completely effective solution to the problem, all without any nerfs.

Group Ambush becomes a modifier which is applied a bit differently. IE; Group Ambush is instead a value that can be up to 560 (8x70). This is divided equally among all units and rounded down. So in a stack of 8, you have your +70% bonus per unit, and in a stack of 24, you have your individual bonus of +23% each. Statistically speaking, you'll still be better off with more units in most situations. However, this will truly reflect Ambush, as travelling with less units would lend you a stronger per-unit ambush ability! Smaller crew, easier travel/stealth/ambush, but less overall strength and security for your hero and/or ghosts.

Note that this is a simple and perfectly effective solution. Assassins won't even be totally nerfed on defense, as the bonus will climb after each battle as defenders' numbers are reduced. It also restores the strength of early Assassin heroes vs neutrals without giving them any new advantage. For simplicity, we could have a cap of 600 group ambush points. The cap should remain in place, so that you don't see stacks of 5-6 100% ambush units abusing the smaller=stronger mechanic.

[Moonknight]

Good ideas Argammon and LP!

I like the totaling of Ambush points and then dividing across the stack a lot actually. It's harder to pull of an ambush with 32 people than it is with 4.

[KGB]

The cap has to move be capable of going than 600 total points when dealing with lots of defenders.

The reason is that the cap defines the average number of units that can be lost to group ambush. 600 points = 75% for 8 units. That means that 8 crows can kill on average 6 enemies in a stack from ambush. So even if you spread that 600 points over 32 defenders (19%) it still only means 6 deaths on average to ambush (the smoothing function (90% rule) will prevent extreme outcomes of ambush).

So a high level hero or even a strong stack still will have a decent chance to walk 32 defenders with a total group ambush of only 600 points. In which case it's now not strong enough to stop super stacks.

The max 560 for an 8 stack we have now seems right (70%). I'd suggest adding an extra 20 per unit over 8 to the max so that 16 defenders would have a pool of 720 points, 24 defenders would have a pool of 860 points and so on.

With 16 defenders and 720 points that works out to 45% per unit or 7.2 kills on average from just the ambush. That seems about right in terms of units needed to stop a super stack purely on ambush.

KGB

[piranha]

Sounds like a good idea. I've written it down. Will try to add it by 0.9.

[KGB]

Piranha,

You might want to add +25 per unit over 8. I ran a bunch of simulations and that seems to give pretty reasonable outcomes in terms of stopping super stacks while at the same time not making cities too hard to take.

KGB

[smursh]

How will this work for very small units. If I get this right a lvl one assasin could expect say 64 total points(equil to his current value for a stack of eight). If I then send him out alone does he get 64 plus his personal ambush of 20%?
Or does this calculation only apply when the stack gets over 8(or perhaps a max at the calculation value for 4 units say, in this case it would be 32) Otherwise there is no point to the barbarian.

Beyond that I think this is the best suggestion ever for dealing with the ambush problem.

[KGB]

Smursh

It only comes into play in cities where you can stack more than 8 units. In a normal stack of 8 you are limited to 70% per unit as it is now.

Individual ambush (Orc, Wolfrider, Medusa, Ghost, Assassin hero) still gets added on top of that as it is now.

KGB

[LPhillips]

Piranha,

You might want to add +25 per unit over 8. I ran a bunch of simulations and that seems to give pretty reasonable outcomes in terms of stopping super stacks while at the same time not making cities too hard to take.

KGB

KGB,
I'm afraid that could defeat the purpose of the "nerf". It's to make Ambush no longer a perfect city defense, but relegate it instead to interception without making it completely useless in defense. It will be the best and only feasible city defense again if you need many, many super stacks to kill a city with one bonus in it. Because: Ambush isn't useful attacking a city, and if both players have roughly equivalent units the defender will always enforce a stalemate as now.

Also, ghosts will need a cost increase. At +64 points of group bonus each, they are too inexpensive. I was pondering this last night separately from the point system, but: ghosts are by far the best bargain, being that they move quickly, need 3 turns to produce, and fly, so are an unrivaled interception force. This was not a problem previously because 1) people didn't know how effective Ambush was, and 2) the amount of purchasing power available to players has been constantly increasing.

Edit: I know you like specific illustrations, so here's one for you: Each side has dragon production and griffon production, and each side has ghost production and a moderately leveled Assassin. One Assassin takes up residence in a city and forms the unbreakable god defense. Let's say on Crescendo, a map designed around choke points. He throws in a bonus unit or two and some fodder. Now, the other player must expend massive resources to overcome the city, meaning he'll sit on a stalemate. Even siege weapons like catapults won't contribute much to the fight.

With the original suggestion, the super stack stopping will happen naturally. A maxed bonus of 560 will kill a super stack, if there is anything beyond fodder units present (and if there isn't, the super stack deserves a chance to win). With many defenders, a synergy of ambush and organic bonuses will defeat a few stacks, with the power of the ambush growing as the number of units shrinks. That's plenty of strength for Ambush on defense. This really means that sending in some strong stacks to weaken the defenses by reducing the number of units will do nothing to mitigate the effect of Ambush on defense.

LP

[piranha]

I've sometimes been thinking about an area based bonus. It could be an idea for group ambush. The idea would be that a group bonus would be on the unit itself and then 15% less on unit next to it and dropping by 15% per unit away. The fight order would control if the bonus comes into play. For ambush in a city you could not place your assassin at the far back of 32 units because the front units would not receive the benefit from it.

[KGB]

LPhillips,

I'm afraid that could defeat the purpose of the "nerf". It's to make Ambush no longer a perfect city defense, but relegate it instead to interception without making it completely useless in defense. It will be the best and only feasible city defense again if you need many, many super stacks to kill a city with one bonus in it. Because: Ambush isn't useful attacking a city, and if both players have roughly equivalent units the defender will always enforce a stalemate as now.

Again, it's not Ambush that's the problem. The problem is that cities spread bonus's over 32 defenders. That's ultimately what's going to have to be fixed in the game. Otherwise we'll be back here discussing why 27 HI boosted by a Unicorn/Devil/RD/DK/Pally are unstoppable. I mean even a L5 DK with -9 Chaos, a Unicorn and a Devil and L6 walls and HI will take crazy amounts of units to kill simply because of how much effect those bonus units have over 32 defenders.

If you cap the max group ambush in a city at 560 then the example simulation I gave (Hv Inf, 1 hero) won't be capable of stopping a super stack PURELY ON AMBUSH. 32 units with 10% ambush won't kill any more than 8 units with 70 because the smoothing function won't allow it (in other words 3 Orcs with 8% ambush will never ever get 3 ambush hits because that's far outside the norm. They won't even get 2 ambush hits). In that case Ambush won't be useful at all because a super hero stack will rub his hand with glee at a city with 32 weak defenders limited to 560 pts of Group Ambush because it's just a gob of free XP.

If it's *that* much of a problem the game is better off changing Orcs to 15% Ambush, WolfRiders to 25%, Ghosts to 40% and drop the Group Ambush power entirely from the Ghost. Then adjust the Assassin to 5% Gp Ambush per level capped at 25% and increase his Chaos in cities to +2 for 10 points so he has something else to invest in.

But instead I'd rather fix the problem of cities spreading bonus's over 32 defenders because we'll be back here again nerfing another power before you know it.



Piranha,

Your suggestion might be OK. But it needs to work for all the other bonus's too (Morale, Fear, Chaos, Leadership, Negates) because those have the exact same problem. 1 unit spreads it's powers over 32 instead of 8.

There are some obvious things that could be done:

1) You could divide the bonus's by the defending stacks. In other words if you have only 1 RD in a city and 4 stacks then each stack would only get 2.5 morale instead of 10. So that it would require 4 RD's in a city (1 per stack) to get the full 10 morale. Or maybe the stack with the RD gets full 10 and the others only 1/4 rounded down or 2. That would mean Ghosts would only give 2% Gp Ambush to other stacks in a city instead of 8 and Medsusas would only give 1 Fear to other stacks, Devils Negate would only be 2 for other stacks etc. Makes combat display a bit more complex but it really helps with 1 unit spreading too much power. Needs a bit more fleshing out obviously.

2) Reduce the number of stacks that can be in a city. Right now it's always 4. But what if the number of stacks allowed was tied to the wall level?
Wall Level 0-2: 1 stack
Wall Level 3-5: 2 stacks
Wall Level 6-7: 3 stacks
Wall Level 8: 4 stacks
Now you need to spend a fair bit of gold to put more stacks in the city. On the map you can just place a Red X over the other squares to indicate they can't hold stacks (Ideally graphics would change to reflect city levels but that's going to be costly). You'd still allow movement through those squares at a cost of 1 so armies can enter/exit but just can't stop on those squares until the walls were upgraded.

3) Or you can introduce new powers similar to what DLR had in Poison (-10 combat)/Disease(-1 hits). Basically powers that last a full turn so that multiple attacking stacks can benefit from them. So an attacker could poison/disease most of a city with 1 stack then use follow up stacks to beat the weakened defenders.

KGB