Uncommonly Good: A Story of Elves

[Mardoc]

It, um, appears you need something exciting, or at least noteworthy. Since I would prefer that excitement be directed other place than lairs, let's talk about Tupperware!




What? What's wrong with Tupperware?

Ok, fine. Let's talk about Fire in the forest!



As you may be aware, there are 4 main ways for jungles and forests to catch fire in this game (technically, creates a Smoke).

1. A Fire adept sets the fire
   
2. A Pyre Zombie explodes
   
3. The Sheaim cast their world spell
   
4. A ritualist casts Ring of Flames
   

Coincidentally enough, Nicolae can (well, eventually) do all of the above! We can do 1 and 4, and probably many/most of our opponents are in the same boat.

'Burning down the elven forest' is often discussed as an easy way to threaten the Ljosalfar. How realistic is that, anyway? I decided to make a worldbuilder test and play around with it.

But first, some info from the XML

• 3% chance/turn that Smoke becomes Flames
  
• 20% chance/turn that Flames spawn a smoke in a surrounding tile
  
• 20% chance/turn that Flames turn into a Burnt forest
  
• 3% chance/turn that Burnt Forest turns into a New Forest
  
• 3% chance/turn that New Forest turns into a Forest
  
• 3% chance/turn that Forest turns into Ancient Forest (when FoL)
  

What's striking here? Well - first and foremost, I couldn't find any code that makes Smoke go away - only Spring seems to do that. It appears to stick around forever, or at least until it becomes Flames. Once it's Flames, we're talking ~5 turns to go out and become Burnt Forest, and then ~33 turns/step of upgrade back to where we want it. Burnt forest cannot catch fire again, nor can New Forest, but the essential random nature of the upgrades means it's quite reasonable for a tile to grow to Forest while a Smoke still smolders on a nearby plot.

While a terrain is Smoked, it's a -10% defense. Flames, of course, are definitely impassible to anyone but Fire resistant creatures - mostly demons and orcs. I believe, but am not certain, that Smoke also causes unhealthiness in neighboring cities.

How much of this matches a world builder test? After all, it's possible that the relevant code is in a different file than those I looked at, or a differently named variable, or some such. Let's find out!

I made a worldbuilder file, and set 4 Smokes going, and waited 50 turns. What happened?

Well, the first two just smouldered:


Yep. 50 turns from casting Blaze, and there's still nothing happening. This isn't terribly surprising, assuming I read the XML correctly. There's a 22% chance for a tile to make it 50 turns without any change.

Next, we had a couple examples of fires that spread:

This one burned 7 tiles in 50 turns, and created another 15 Smokes. Annoying, true. Smoke is -10% tile defense, so it would hurt our forest fortresses, and if we make it to Ancient Forests, well, that's 7 tiles that have to start over from scratch. And if we let it burn another 50 turns, probably many of those Smokes would also go to Flames, etc.

It turns out that it's possible for a tile to turn to flames and then go out in the same turn, leaving behind a Burnt Forest. This is probably the best case, that it only wrecks the tile it's on.

Also, notice the pattern - when Smoke spreads, it spreads to *every* adjacent tile to the flames, rather than checking each tile individually as I expected.



This one burned 9 tiles, but only has 8 remaining Smokes; there's a chance it might burn itself out, although admittedly I didn't test this yet. Also - notice that some of the Burnt Forests have already become New Forests.

So, what are some general lessons to take from all this?

• First, Fire is inexorable. If you don't do something, it will burn until there is nothing unburned. No Smoke will go out short of at least burning the tile it's on, and it's likely to set Smokes in all the adjacent tiles as well. The only exceptions are luck, or a firebreak. Unless you're willing to deal with a perpetual fire, or chop down forests to be firebreaks, we need Spring sooner or later.
  
• Second, surprisingly, Fire is pretty slow. 20-30 turns/tile forward! Spring therefore is not an urgent requirement. We also probably don't need a huge number of Water adepts - if you can find the Grigori or Elohim soon enough and get a 10T loan, that's probably enough (so long as we protect the relevant adepts).
  
• Third, as part of the nature of dealing with low probability events, there's a lot of variance. Fire ends up being quite unpredictable; as you saw, we had everything from nothing happening to enough fire to really mess up a city.
  

To conclude? Spring adepts are not a 'nice-to-have', they're a requirement. Someone, sooner or later, is going to make our forests burn, and it's Spring or suffer then. On the other hand, they're not an urgent need, either. Even when the fires do start, they're slooooww.

If we can get some fires set in the jungle (say, by killing Pyre Zombies in Sheaim lands), odds are good of both it eventually turning all the jungle to forest, and also taking long enough that we can walk, not run, to Spring adepts to control the edge of the fire.

And we definitely do not need to worry about getting spring adepts as part of any T100 force. So long as the fires aren't starting in our lands (and with Hawks and PoL, we ought to be able to avoid that), we'll have plenty of time to deal with them eventually - 50 turns to transform 1 Smoke into 7 Burnt Forests is not a rate that scares me. By T150, say, we'd probably want to have one or two Spring adepts.

[Ravus Sol]

This was all tested with a single starting plot yes?

Not from a adept casting blaze everyturn along the borderline or pyre zombies exploding everywhere? :P

[Man Behind the Mask]

A TURN! The results... meh

Score increases: Us (28, 3rd ring land points) Kyan 6 (pop) Grigori 6 (pop) Nicolae 12 (3rd ring land for a coastal capital?)

I’d written a long thing about 12 points being a really odd number for Nicolae’s score increase, and then went back to look at screenshots. 114-103 is 11, so we’re actually looking at a simple tech. It’s only six turns since the last tech (assuming he hasn’t got lucky with a hut), so it seems likely he’s gone for something other than Bronze Working next. We’d probably have to check out his GNP graph to consider the possibilities, but I’d say agriculture is very likely. Not sure if that diversion away from flaming corpses is good or bad.

Gilden Investigations:

Thanks for doing a much more comprehensive job than I did, Gaspar, and thanks to Azoth for remembering the flaw in our excitement.

On the game mechanics, Defensive Strikes only work when the unit is not defending itself – what makes the difference when Gilden is top of the stack is the 4-7 First Strikes taking chunks out of an attacker before getting to normal combat rounds. I’m still a little puzzled by Blitz seemingly granting multiple Defensive Strikes on the same attack, as I’d thought it was just that you could perform multiple Defensive Strikes in the same turn. Really is a shame we can’t see the combat log in this kind of situation!

My test was on flatland with Hunters, so it’s no surprise archer Gilden saw a lot more action in a city assault. I agree with Mardoc’s general conclusions before he saw the actual results – PZ fire collateral means that any target is at serious risk whether we have a highly promoted Gilden helping out or not. I have some experience with Gilden against AI Sheaim, and his big benefit was as the “first man over the top” to get better odds for the next attackers. It will be interesting to see if we can squeeze in Archery and get Gilden built a little while before the deadline, so we can get some value out of this analysis.

One risky possibility worth considering is to split our forces. We could rely on an overwhelming treant surge to wipe out his main stack coming up from our south. Then an attack force of Priests and Hunters could move in via the east to find his homeland mostly empty. This could potentially go disastrously wrong, but then so could slamming into a PZ stack with anything other than highly superior numbers.

Flaming Trees:

I’m pleasantly surprised that a fire is quite unlikely to start from a smoke. It actually suggests “blazing” the jungle could be a very slow process if you have to wait for the fires to start. The coinflip of burning out versus spreading to all the adjacent tiles is a lot to leave to the RNG, sadly.

Getting our hands on water mana in a trade is definitely the best way to go. Elementalism is a fair way down the Sage bulb list, so we’d have to research it and use up our single mana node if we need to rush for Spring.

[Mist]

On the game mechanics, Defensive Strikes only work when the unit is not defending itself

I'm sure I had Gilden launching them as an one man army. You might want to WB test this.

[Man Behind the Mask]

I’m pretty confident, but I am Gaspar's noob assistant rather than his vet source of the truth. PB covers his understanding here. Thoth answers a question in the PBEM1 lurker thread here. Logs in my post earlier show a defensive strike when one of the Hunters is being attacked but never when Gilden is the top defender.

In FFH Adventure3, my super-Werewolf has had Drill IV for a while and been the top defender 20+ times in a row as the AI throws units at a stack. He’s never made the defensive strike when defending from what I can recall. [Yes… I am still slowly playing the Adventure and got probably my favourite screenshot ever when said Werewolf streaked halfway across the empire to obliterate Stephanos ]

[Mardoc]

This was all tested with a single starting plot yes?

Not from a adept casting blaze everyturn along the borderline or pyre zombies exploding everywhere? :P

Yes - but then both the fires expanded to a full 3x3 square by T16, and they still didn't burn a heck of a lot, or spread very quickly. The northeastern fire made it three tiles from that smoke ring, in a random direction, in 34 turns; the southeastern one only made it 1 more tile in any direction in that time.

I can run the test again, (or you could, with my file) starting off by creating a line of smokes, but I don't expect the answer to be much different - inexorable, maybe slightly faster, but still slow compared to, say, the Pyre Zombies themselves. And therefore more of an annoyance than an existential threat.

[NobleHelium]

I’m pretty confident, but I am Gaspar's noob assistant rather than his vet source of the truth. PB covers his understanding here. Thoth answers a question in the PBEM1 lurker thread here. Logs in my post earlier show a defensive strike when one of the Hunters is being attacked but never when Gilden is the top defender.

In FFH Adventure3, my super-Werewolf has had Drill IV for a while and been the top defender 20+ times in a row as the AI throws units at a stack. He’s never made the defensive strike when defending from what I can recall. [Yes… I am still slowly playing the Adventure and got probably my favourite screenshot ever when said Werewolf streaked halfway across the empire to obliterate Stephanos ]

I'm pretty sure defensive strikes work if the unit is the defender. Your werewolf probably used its attack during your turn and doesn't have Blitz.

[Gaspar]

Let's see, lots to discuss here...

On Gilden:

In my tests, Gilden definitely did Defensive Strikes when he was both top defender and not top defender. The lone instance of him doing multiples on one round I seem to remember him as top defender, but I didn't log it so could be wrong. Blitz is not available to us anytime soon, but Drill IV is what's making the D4 + Blitz combination work so well, not Blitz. The decision is really a matter if we want to make him an excellent defender down the line or if we want him to be a versatile attacker. I don't see us getting Archery significantly before the zombie attack, because ultimately, we need mobile units more than we need stronger ones. The test, if it showed nothing, showed you have to hit PZs out in the open. Hawks and Hunters/PoL/Treants from the worldspell are going to be key to that. And shortly I'll make the great planning post with regards to that. Really, anything that we can't get 2 moves out of is just not going to help, the splash damage from the zombies is too strong otherwise. Fyrdwell would actually be quite useful here, but we can't get HBR in time either.

On Fires in the Forest:

First off, thanks for the research, Mardoc. And the Smokey the Bear reference. I think Smokey had it right, really. Only I can prevent forest fires. Its going to have to be a combination of having the inevitable Sheaim war on Nicolae's turf and then good diplomacy to keep the others from Blazing in our lands. Hmm. Given our civ's theme, shouldn't we encourage blazing? The best way out of the fires is the constellation event that puts out all the fires in your lands. Since you can't count on that, we probably will eventually need spring adepts. Its worth noting, however, that nobody has really aggressively targeted Ljosalfar forests in other games. Not saying we can count on that, but I don't think we want to turn our lone node into Water either. A trade with Brian or mackoti/Acostalas is probably the best idea. Downside is our palace mana is pretty uninspiring for trades, especially if nobody else has a reason to fear the Sheaim and their undead army.

Other stuff:

Yeah, clearly an 11 pt increase for the Sheaim there. I'd bet solidly on Agriculture. Patently foolish to go Mining before Agriculture, IMO, unless you have a gems resource or something at the capital. But I'd guess he'd have gone BW next if we hadn't NAP'ed him. Regardless, its a safe bet he'll have at least Bronze PZs at the war.

[Man Behind the Mask]

In my tests, Gilden definitely did Defensive Strikes when he was both top defender and not top defender.

I stand corrected, surely for neither the first or last time . My own tests in game confirm that Gilden definitely does make defensive strikes when he is also the defender.

Doing the checking reminded me that he is eligible for the "Keen Sight" promotion, which ups his defensive strike probability to 90%. This might be the best use for a late-game Gilden where his lower basic strength means he won't be that valuable on offense. Everything will depend on when we get him and how much XP we have to play with when he needs to be used in earnest.

A trade with Brian or mackoti/Acostalas is probably the best idea. Downside is our palace mana is pretty uninspiring for trades, especially if nobody else has a reason to fear the Sheaim and their undead army.

There is another mana source nearby, which tends to be much more useful .... shame it's buried in the jungle.

Patently foolish to go Mining before Agriculture, IMO, unless you have a gems resource or something at the capital.

I guess he can guarantee settling any available copper with his second city?

[uberfish]

Yeah I tried to burn jungle with ritualists in PBEM2 to convert it to forest. It takes forever.

Also clearly you guys need to explore more lairs instead of spamming the thread so much.