Home

  • Avoiding things because I’m avoiding them

    This is going to be a short post. I HOPE to have a better one up this weekend.

    But to cut to the chase, sometimes I avoid things… simply because I started avoiding them previously and now I feel too guilty to just fess up and DO them. 

    I get an email in my inbox that I don’t want to look at, so I ignore it for a day. Then the next day I feel guilty, “what will they think of me, that I was ignoring them for a day,” so I keep ignoring it. Obviously ignoring it for two days makes me feel even more guilty, so this is a problem that quickly spirals into me just ghosting someone for weeks until I finally write a long apology and actually just respond like I should have.

    And it happens with this blog too. My schedule has slipped, I wanted to write a post every weekend, but now that I’ve missed a couple I suddenly feel very guilty, and that just makes me want to avoid doing a post even more.

    I need to overcome these feelings, and I’m sure everyone has them. But to day, I still don’t know how. Having someone else with me when I read emails or write at least puts some of the sting off, it reminds me I’m not alone and gives me someone to bounce ideas off of. But people can’t be with me all the time, this is a skill I need to have for myself.

  • Is it culture? Or is it incentives?

    The Internet in general is US-centric. So even on the European parts of the Internet it’s common for countries (or the entire continent) to compare themselves to America. There are thousands of things you could compare, but the most contentious is probably the economic comparisons. America has recently grown much more strongly than Europe, and it doesn’t take an economist to realize that nearly all of the world’s top companies and startups are located in America. San Fransisco alone has more billion-dollar startups than entire countries, and before you say “that’s just silicon valley,” New York and Boston aren’t far behind.

    There are a million ways to explain this discrepancy and plenty of reasons why Europeans may even think it’s good. We could talk all day about whether worker’s rights are fundamentally incompatible with cut-throat capitalism, and if Europe has therefore chosen the better path. But the most flawed reason I see bandied about is that Europe just has the wrong culture for this kind of stuff.

    Europe is more laid back, less aggressive. Their investors prefer same, consistent gains. The European mindset isn’t focused on innovation, and culturally Europeans aren’t focused on business the way Americans are.

    I think these explanations are wrong and dumb, and I’d use more expletive words if I hadn’t made a New Year’s Resolution not to do so in my writing. I don’t think Europeans are culturally less attuned to startups and Big Business, I think the legal framework prevents it.

    Not long ago, Europe was seen as the beating heart of innovation and technology. Industrial progress, scientific progress, just go to any chemistry or physics class and see how many formulas are named for Germans. But now America dominates the industries, and I think it’s because of government, not culture.

    The American business framework provides significant bankruptcy protection. People mocked Trump for his many bankruptcies, but most investors know that 90% of good ideas fail and the last 10% have to cover those loses. Bankruptcy is a way for investors to mitigate their downside, and thus allows for bigger risks to be taken.

    The American financial system also gives significant benefits to investors, giving them greater flexibility in buying and selling their company to whomever they wish. Until Biden and Trump brought protectionism back to the fore, it was not uncommon to see American companies sold to foreign investors with little fanfare. Nativists and racists may complain about *gasp* Chinese people owning an American company, but from the investor’s perspective selling the company is a good way to cash out his winnings from the investment. Foreign buyers compete with American buyers, and this increase in demand means prices go up. This means the sale price of companies goes up, and that increases the returns on an investor’s investment.

    But long before Trump, Europe was made famous in the tech world for blocking foreign buyers from its companies. Again, nativists wrongly think that this strengthens the European tech industry by “keeping it in European hands.” But when an investor sells out, they get cash in return. What do you think they do with that cash? They don’t hoard it like Smaug the Dragon, they reinvest it. Because they’re investors. By blocking foreign buyers, you reduce buying pressure, you reduce how much money investors can get out of their investment, and you therefore reduce their upside potential. Is it any wonder then they’d prefer a safer investment, when Europe is happy to cap the gains on any risky tech investment they make?

    And Europe prides itself on fining big tech companies for any reason whatsoever. But surely it’s obvious that a government hostile to profitable tech companies would scare off anyone wanting to make a profitable tech company near them. Better to start in America or get out of Europe ASAP. Microsoft and Apple can afford billion dollar fines, but such sanctions could be lethal to a smaller European tech company. So again investors are scared off, entrepreneurs are scared off, and Europe wonders why it doesn’t have a tech sector.

    “But what about ASML and Spotify!” And what about them? For every single, solitary European company that manages to rise above the hostile governing environment, there are 10 American companies that rose under easier circumstances. Spotify started in 2006, and since then Massachusetts alone has started Draft Kings, Moderna and Intellia Therapeutics, all of comparable value to Spotify. And Massachusetts has half the population of Sweden.

    People respond to incentives, and the incentives for risky tech investment are very poor in Europe. Bankruptcy is easier in America, returns are (or were before Biden and Trump) less likely to be capped by protectionist policies, and (before Biden) the government generally has taken a more lax approach to dealing with corporations. You can debate if these things are good or bad, but I find them far more likely reasons for America’s tech dominance than “culture” or “attitude.”

  • Dominions 6: Out Now

    I wrote earlier about Dominions 5 and its many complexities. It’s an incredibly deep game with a lot of moving parts, from sacred troops to magic research to recruiting or summoning mages. And I even outlined some of my favorite strategies in later posts.

    If you enjoyed those posts or were interested in trying Dominions for yourself, Dominions 6 is out now. This latest installment brings about 5 new nations to the title, raising to total to I think somewhere north of 80. And each nation is a wholly unique beast so it’s really fun to craft perfect, elaborate strategies for each one. 

    As it’s just come out, the multiplayer community is at its most active, so now’s the best time to play multiplayer as well. I’ll still be thinking about the game more than playing it, but if you buy it too I hope you’ll realize why even just thinking about it can be very fun.

  • Doing the Possible: When is it Impossible?

    I recently wrote about “If We Can Put a Man on the Moon,” the book that wants to teach people how to do government well. Some of their message is simple: success in government requires a good plan executed well. But while they want their message to be non-partisan and universalist, I’m not sure it can ever work that way.

    The big question I have is this: when is failure because of a good plan done poorly, and when is it because of an impossible plan that would never succeed? For instance, the book lays plenty of criticism at Nixon’s price controls and Ford’s purposeless “WIN” buttons, and it does so by saying that price controls and government nudging cannot control inflation. The book agrees with Milton Friedman than inflation is a monetary phenomenon, solved by Volcker when he hiked interest rates.

    On the other hand, the book criticized many plans for their implementation rather than their ideas. Boston’s failed bussing experiment of the 70s is excoriated for how it was done with no real plan or input from the community. But is bussing ever a good policy for implementing desegregation? Many have looked back and said that no, bussing was never going to work. It was unpopular amongst both white and black communities. Just look at this blast from the past:

    A majority of Americans continue to favor public school integration, but few people—black or white—think that busing is the best way to achieve that goal, the Gallup Poll reported yesterday.

    Five per cent of the people in a recent survey by the organization—9 per cent of the blacks and 4 per cent of the whites—chose busing children from one, district to another rather than several other alternatives.

    New York Times

    Most of those interviewed preferred either changing school boundaries or providing low-income housing in middle-income neighborhoods as preferable plans for school integration. 

    In the same vein, the book knocks the Iraqi occupation for having no plan for creating a stable, post-Saddam Iraq. But was that kind of “nation building” even possible for the US military to achieve? Especially in a country with such vast cultural and ideological differences to ours? 

    I remember going to school with a guy who served in Iraq. He talked about how he was tasked with keeping Iraqis safe by removing weapons and disarming citizen. He once came to the tent of a Bedouin he thought had a gun and ammo. And when the Bedouin refused to let him search the tent, he simply ordered his troops to cut open all the Bedouin’s bags of rice, ruining his food but finding a hidden AK-47.

    The soldier then said that he told this Bedouin “look, you should have just made this easy for us,” but all I could think of was “wow, this is why they fucking hate us.” This soldier just proudly violated the rights that we in America would call the 2nd and 4th amendment, and if he’d done that in America it would be a national scandal. Iraq may not have our constitution, but they still probably feel entitled to basic human rights of dignity and property. Even if we amended our constitution to remove the 2nd and 4th amendments, how would any American feel about armed military personnel breaking into their house, upturning all their belongings, and then stealing their stuff? 

    So was “nation-building” even possible? Or was it, like bussing, an idea doomed from the start?

    This is the difficulty in analyzing good governance, by focusing on the process you implicitly assume the idea is workable. Now, the authors do mention some ideas that they find impossible. They chide Nixon’s price caps because price caps can’t fix inflation, which is a monetary phenomenon. I happen to agree with them, but that’s because both I and the authors ascribe to an orthodox economic framework. A socialist would disagree with us, saying price caps are perfectly valid but that Nixon just used them poorly.

    So a socialist might see Nixon’s price caps as a failure of implementation and not a failure of ideology. And while the authors see bussing and nation-building as failures of implementation and not ideology, a school choice advocate and a non-interventionist would disagree and say that for those ideas, a successful outcome was never possible. So how do you judge policies by their process, when people can’t agree on their possibility?

    Ultimately, I think “a bad plan” vs “a good plan, poorly executed” is a political question for which there is no agreed upon answer. And to that, while the tenants of the book may be accepted broadly, it won’t do much to change the tenor of governance even if everyone in America agreed with it. All of politics is about the disagreement over “which plan is good,” and “how do we execute a plan well.” So telling people to “have good plans” and “execute them well” is sort of like telling a sprinter to “just run faster.” It’s advice that does nothing.

    I think the book is good, I think it’s well worth a read by anyone interested in politics. I just think it’s impact will not be too great even in the minds of its readers.

  • Doing the Possible: Musings on good governance

    I’ve been reading “If We Can Put a Man on the Moon,” which is a book that attempts to explain why some government policy succeeds and some fails. The book outlines how public policy requires a clear objective, a clear plan to reach that objective, and the ability to follow through with it. It all seems rather obvious when you write it out like that, but the book offers some definite insights.

    A clear objective seems obvious, but is surprisingly easy to overlook. Gerald Ford promised to “whip inflation now,” but how exactly did he expect to do that? Supposedly the government would politely encourage citizens to do things like grow more food and use less fuel, to increase supply and decrease demand. It’s a nice idea, but polite encouragement doesn’t move the economy, and Fords “WIN” policy went nowhere.

    A clear plan is also something that seems obvious, but often gets overlooked. When California reformed its electric grid in the 90s, no one had really thought through how the new system would work. They set mandates to ensure that prices were capped for consumers, but did nothing to ensure adequate supply. It was legal, for example, to buy power at a low price in California and export it for the uncapped price in other states. Then, if California didn’t have enough power, the utility was obligated to import power from other states no matter the cost, but was not allowed to pass this cost on to customers. This lead to companies easily gaming the system by exporting power for cheap, then re-importing it at a higher price. 

    People like to blame greedy companies for the failed California power experiment, but companies are always greedy in all cases. The government should create a system in which corporate greed leads to societal good, such as how tech companies have given us ever better computers at lower and lower cost. Failure to plan leads to a system that is designed to fail.

    The ability to follow through is a common complaint, but it too has unexpected pitfalls. The political class has different incentives than both the bureaucrats and the people, but they all work together for a plan to succeed. Politicians have an incentive to pass a bill and say they “fixed” something, that’s why most celebration happens on when a bill is pass instead of 5 years later when its effects are being evaluated. Bureaucrats are just career workers like anyone, and have an incentive to do their job and get paid. They aren’t incentivized to go above and beyond for the benefit of a politician who might not be there in four years. 

    This is why it’s so common for politicians to take office promising “big changes” but still not accomplish much. Once the bill is passed and the photo-op is finished, it’s out of their hands and they don’t has a reason to keep caring. And when someone comes in saying they’ll “upend the stuffy bureaucracy,” well if they don’t meet the career employees halfway they’ll engender resentment in a group that can drag its feet and wait for the political will to die down.

    All told, the book does have a lot of prescriptions for good governance:

    • Have an idea for how to fix a problem, and don’t make a plan of action without a strong idea. Likewise, seek out good ideas from everywhere, and be willing to challenge your own ideas to see if they’re actually appropriate.
    • Make a rational design for how the problem will be fixed. Focus on the design, not just on getting buy-in from the right pressure groups. Stress-test the design and hire people to poke holes in it. Then fix those design holes before passing a new law.
    • Ensure oversight and continued evaluation even after a law is passed. The job doesn’t end after the vote and the signature.
    • Understand that government is different than any other sector, and that you have to meet people halfway. You can’t treat public employees or the public at large as workers in your company or as cogs in a machine. 
    • Don’t assume the success of a plan. And don’t assume that just because it hasn’t failed yet that it won’t in the future. Look for any signs that cracks are forming, and fix them before they get too big. The space shuttle Colombia flew 27 missions, many of which showed problem signs, before the fateful 28th mission that ended in disaster.
    • Keep re-evaluating. If a program is no longer fit for purpose, fix it, replace it, or kill it. Be willing to see that something isn’t working and be willing to change it. And don’t keep trying the same program over and over without change, be willing to go back to the beginning and look for new ideas and new designs.

    That, in a nutshell is what the book is about (or at least my reading of it). It’s certainly more uplifting that what you expect from a book about governance, but without ignoring the data and the details. I’ll have more to say on it later, but I think anyone who likes this sort of thing should check it out.

  • My kingdom for a venv

    I’ve never enjoyed using Python. I think my feelings on it can be summed up by this video. But for whatever reason, Python is unavoidable if you want to do anything with AI/machine learning. And so as someone wanting to get into AI, I have no choice but to use it.

    But I don’t have to learn to code it of course, because all the tools you need for AI area already written and available. ChatGPT is of course easy to use on the web. But what if you wanted to have a version of ChatGPT that was snarkier, or wrote better jokes, or was in whatever way tuned specifically for your needs and wants? In that case, you can always make a fine-tuned language model and use it yourself.

    But that’s where Python rears its ugly head. I wanted to fine tune a language model. So I installed LLaMA, downloaded a simple model from huggingface, and got to work. 

    To fine-tune a model for your own needs, you need to have data and you need to annotate that data. No time to explain how annotations work, but there are programs that make it easy. There is a program called Label Studio that I thought I could use. The instruction say to just download python, make a venv (virtual environment) and have pip (a python installer) install Label Studio. Sounds easy, right? Just 3 lines of code.

    The trouble started almost immediately because despite Label Studio telling me it was available for Windows, the install instructions were actually written for Linux. I realized this and corrected it, but the trouble didn’t stop. Once I created the venv, I tried to install Label Studio, but one of the dependencies failed to install so the whole process failed.

    Uh… what? Why is this program, which is available as a paid enterprise product by the way, failing to install itself due to a dependency issue? I find the missing dependency and try installing it directly to the venv, hoping that fixes the issue. But no, it still errors out. What am I missing?

    So it turns out that when I directly install that dependency, it installs the latest version of it. But Label Studio is looking for a specific older version, so it still tries to install the older version when installing itself. I tried to install the specific older version, and that fails too. Apparently I can install the new version with no issues, but not the old version.

    Reading the message closely, it says that to install the old version I need to have another python module installed and also add that other module to the system path. Now we’re getting into part of why I hate venv. The thing about Python is that if you install itself outside of a contained environment, it infects your computer and doesn’t get out. Ask an amateur pythonist how to remove an old version of Python, and see the blank look on their face. Just deleting the folder doesn’t fix it.

    And this old version/new version bs can mess you up something fierce, because some other python module will start looking for what it needs, and find the old version instead of the new version. Or it will be sent to where the old version used to be, but finding nothing there it will error out. Venv is supposed to fix all this so you only install things into designated containers where they can’t escape.

    But I can’t do that, because to install something into this venv, I have to install another package and add it to the path of my entire Windows system. So the venv isn’t even doing what it’s supposed to do!

    So I gave up. I hate having to use python like this, normal programs will just come to you as an executable or a zip and you use them. Python always needs to install itself everywhere and then usually fails even then. So I won’t use Label Studio and will look for another tool instead.

    If anyone knows of a good annotation tool for LLM data, hit me up.

  • Thoughts on the new year

    I don’t have a full post to make, I’ve been lazy over the holidays. But I was thinking that while the New Year is traditionally a time when people commit to change and self-improvement, it doesn’t have to be the only time. You can split the 365 days of the year up however you want, and declare any of them to be the start of a “new year” in which you’ll change yourself for the better.

    Most new year’s resolutions get dropped in the first few weeks, and I’ve got some resolution that I’ll likely not keep to. But even if so, I’ll try to pick each of them back up during the year if I can. I don’t have to wait a full year to only start things on January 1st. I can decide that February 1st is also the start of a new year, and re-engage with my resolutions then.

    I don’t know if this will work, but it’s what I want to try. When I drop off on my resolutions, I want to pick them back up. So I’ll be trying this outlook in 2024.

  • Dominions Strategy: Thunder Striking Wizard Thrasher

    Dominions Strategy: Thunder Striking Wizard Thrasher

    This post named in honor of one of the pretenders from a hilarious Disciples game I watched ages ago. I’m continuing my series of posts about the fun strategies you can use in Dominions 5. I still hope my posts inspire someone else to start playing the game, either now or when Dominions 6 launches in January of 2024.  Last time was physmoss, this time it’s Thunder Strike.

    “Evocation” is the general magic-y word for “big spells that do damage.” Fireballs, shockwaves, if it directly hits someone, it’s usually evocation. I think the separating of magic spells like this first came from D&D, but the tradition has carried on in Dominions.

    In Dominions though, not all Evocation is created equal.  Some spells really aren’t worth it, and casting “retail evocation” aka low level tiny spells is a sign that times are desperate and you don’t have the mage firepower for Big Boy spells. But one of those Big Boy spells, indeed a spell so powerful it people might build their pretender specifically to counter it, is Thunder Strike.

    Let’s start with the good: Thunder Strike deals a boatload of armor negating shock damage to a square, and then a smaller amount of shock damage to all the squares around it. That means any units sitting where the Thunder Strike hits get instantly deleted, no matter how good their armor, while surrounding units can still be stunned by the small shockwave that surrounds the big strike. So this spell not only deals with ultra-powerful enemy units (like the physmoss mage from last post), but also ties down large enemy armies by shocking their units and stunning them for a turn or two. It may not seem like a lot, but stunning some troops can break apart their formation and allow your own army to defeat them in detail.

    The only counter to Thunder Strike is Shock Resistance, which is a very hard resistance to get. This is why some nations will specifically build their pretender around defeating Thunder Strike: if the pretender has high level Air or Earth magic, they can bless their sacred troops with Shock Resistance. That will help immensely against the large stunning shockwave, but still may not be enough to save units from the big Thunder Strike at the center.

    But now here’s the bad news: Thunder Strike requires a mage to be Air 3. Remember that Hordes of Skeletons from before required just Death 2, higher levels of magic are much less common than lower levels. Mages with magic of Level 1 are everywhere. Level 2 is usually doable. Level 3 and above is exceptionally rare. So while it’s easy for your enemies to amass the Death 2 mages needed for skellyspam, it’s hard to find the Air 3 mages needed for a Thunder Strike counter.

    But there is one hope for an aspiring Thunder Striker: and it’s called Storm Power.

    Storm Power is an Air 2 spell that adds +1 to a mage’s Air Magic. In essence, it turns an Air 2 mage into an Air 3. BUT it can only be cast when a storm is already raging.  And creating a storm on the battlefield requires… Air 3 yet again. So Air 2 mages alone can’t make this work, but everything can come together if they can get just a single Air 3 mage.

    The trick is that the Air 3 mage will cast Summon Storm, then the Air 2 mages all cast Storm Power. And now that everyone is an Air 3 mage, they can start blasting out Thunder Strikes like there’s no tomorrow. This is a huge ability, and Thunder Strike plus Storm Power is a key tech level for most air nations in the game like Vanheim and Caelum. 

    The best part is that summoning a Storm will also power up Air Elementals, who Air 3 mages can also summon. Air Elementals can fly directly into the enemy’s lines, barely take damage because they’re Ethereal, AND ignore shock damage. So you can also summon a few of them during the storm and let them tear apart the enemy army while you’re dropping Thunder Strikes on their heads. It’s a brutal strategy.

    This is why Air 3 is such a key breakpoint for mage power. Air 2 is OK, but Air 3 is key. The ability to summon storms and turn everyone into a Thunder Striker overturns a lot of strategies and forces your enemies to come up with effective counters to deal with you. 

    One fun nation I love is Ur: The First City. Ur gets Gudus as a mainstay mage, and 1/2 of Gudus are Air 2. Unfortunately, Ur can’t natively recruit any Air 3 mages, so it seems they’re out of contention for Thunder Striking. But they can use a national spell to summon an Ugalla, which is an Air 3 mage in its own right. So by summoning an Ugalla, Ur can turn itself into a Thunder Striking powerhouse on par with any other.

    And Thunder Strikes have a huge range as well. They can hit well into the back-line and destroy an enemy’s mage corps. So if the enemy sets up skellyspam, Thunder Strike can be a legitimate counter, as you delete their mages to slow the tide of skeletons.

    So that’s yet another tactic I really enjoy. It’s stereotypically done by the “Elven” nations (Vanheim, Helheim those sort) but I’ve also made it work with my favorite nation, Ur. Feel free to try it yourself when Dominions 6 comes out.

  • Dominions Strategy: Becoming Unkillable

    Dominions Strategy: Becoming Unkillable

    I’ve been writing a series of posts about the fun strategies you can use in Dominions 5. I hope my posts inspire someone else to start playing the game, either now or when Dominions 6 launches in January of 2024. Last time was skellyspam, this time it’s physmoss.

    To begin with, lets understand what happens when one unit swings their sword at another. They first roll to see if their attack skill beats the enemy’s defense skill. Then if so, they “hit,” and get to roll for damage. Their Strength + Weapon damage is rolled against the enemy’s Protection, and if they roll higher then they deal damage according to how well they rolled.

    Already we can see the strategy developing, units with high Protection are difficult to harm. But having lots of Strength, powerful weapons, or just an overwhelming number of attacks can still be used to take them down.

    But what is Protection? It’s the sum of your armor value plus whatever “natural protection” the unit may possess, so a being made out of stone will naturally have high “natural protection” even if wearing little, while a squishy human can still get a lot of protection from a suit of plate mail. This is step 1 of “physmoss,” having a very high protection. Mages normally wear robes, but you can craft them a suit of armor. Then they can cast something like “ironskin” on themselves to bump their natural protection into the stratosphere. The sum of that armor plus their ironskin makes for one tough nut to crack.

    But some units have high strength, and crits can roll for high damage anyway. Just having a lot of protection isn’t enough. We can get further by having “physical resistance,” which halves the damage taken from all weapons. This is the “phys” of “physmoss,” and spells like “liquid body” or “temper flesh” can both give physical resistance. Then on top of that we add a bit of regeneration, either with a ring of regeneration or with a nature mage casting a spell. That means after accounting for high protection and physical resistance, what little damage the mage does take, they can regenerate at the end of each combat round.

    But the pièce de résistance is the “mossbody” spell, the “moss” part of “physmoss.” You see, after all the protection and resistances are subtracted out, mossbody then subtracts a flat 15 from any damage taken. In a game where most units will be lucky to deal 15 damage against unarmored enemies, that’s huge.

    So with all this put together, the mage has high protection, physical resistance, mossbody, and regeneration. Their protection is so great that barely anything with every harm them. What little harms them will have its damage halved by physical resistance. And after physical resistance, mossbody will reduce the damage even further. And the tiny slivers of damage still taken are then regenerated by regeneration.

    This combination of spells can make a mage absolutely unkillable. You can send them against hundreds of units and they won’t die. They’ll barely kill, but they definitely won’t die. And eventually even a squishy mage with just a mean left hook can KO enough enemies that the rest get the hint and run off, leaving the mage victorious on the field. A single mage with physmoss can defeat an entire army that forgot to pack a mage of their own who can cast protection-bypassing spells. 

    And that’s what makes the combination so powerful, it’s not completely invulnerability, but it lets a single mage take on armies, forcing the enemy to bring all their resources together if they want to take the mage down. It’s a hilarious tactic when it works and is really fun to boot. So try it yourself if you ever give Dominions a go.

  • Dominions Strategy: Hordes of Skeletons

    Dominions Strategy: Hordes of Skeletons

    I recently wrote a post about how I’m always thinking about Dominions 5 but I’m never playing it. In writing about why I like it, I realize that a lot of it comes down to the sheer number of cool strategies to employ. But it’s hard to get across the huge depth of strategy in a single post, so why not do so in multiple posts? In the next few posts, I’ll give you a taste of all the cool and awesome things you can do in Dominions, and if you think it’s interesting you can do some of them yourself when Dominions 6 comes out in January.

    In Dominions, armies clash in great battles. The army with better weapons, better discipline, or better strength may win, unless the other army has bigger numbers on their side. And what’s a bigger number than infinity? That’s the idea behind skelly-spam, have your mages raise an infinite horde of undead to fight for you, overwhelming the enemy in sheer volume until even the strongest soldiers are ground down by weight of numbers.

    There’s a lot of ways to do skelly-spam though, it isn’t just a button you can press to win. The mainstay skelly-spam spell is “Horde of Skeletons” which can be cast by a Death 2 mage after you’ve researched Enchantment Level 5. So to use skelly-spam, you need a nation with lots of access to Death 2 mages, Death 1 won’t cut it. Death 3 is also good, because higher level mages cast the spell using less fatigue. If mages are casting constantly, they’ll reach 100 fatigue and fall unconscious, no longer casting until their fatigue returns to 99 or below. But a Death 3 mage can cast a lot more “Horde of Skeletons” spells than a Death 2 before reaching that point.

    So to TRULY overwhelm the enemy in skeletons, you need lots of high level death mages to cast it non stop. Some nations can easily recruit lots of death mages, but others may have trouble. Fortunately, there’s a second option.

    Nations with ok death access but lots of astral or blood can also use communions to level-up their skelly-spam. When mages form a communion, the Communion Masters cast spells more easily and transfer the fatigue to the Communion Slaves. Those slaves don’t cast anything, but do regenerate fatigue. However, if their fatigue goes above 200, they start taking damage and quickly die. So in a communion, the masters can keep spamming out skeletons so long as the slaves stay below 200 fatigue. This lets death mages unleash even MORE skeletons than they otherwise could.

    But it doesn’t stop there, because you can level up your communions into “turbo communions.” When a slave’s fatigue is above 200, they take damage, but what if they could regenerate that damage? Then the masters could keep casting for even more skeletons as long as the damage to the slaves is less than their regeneration. 

    Jotunheim is the poster-child for turbo communions. They have “Skrattir” (plural of Skratti) who naturally regenerate 1/10 of their massive HP each turn. They can then have their Gygjas be the communions masters, while the Skrattir are communion slaves. Not only that, one of the Gygjas can cast “personal regeneration” on themselves, and that benefit will transfer to the slaves as well. Now the Skrattir regenerate 2/10 of their HP per turn. Now the Gygjas can cast “Horde of Skeletons” until the end of time, safe in the knowledge that the Skrattir can tank the damage.

    The battlefield effects of this are awesome. Most mages will use their power to buff up (increase the power of) their own troops, then quickly fall unconscious after a few spells. The Gygjas are meanwhile raising an army of the undead. Then, the two armies will meet each other, Jotunheim with an army of the living plus the dead, and the enemy with their army of the living. Jotunheim may be ground down by the enemy’s superior power, but the Gygjas will still be raising the undead. For every Jotunheim soldier that falls, 2 more skeletons will take its place. Eventually, the enemy army will be overwhelmed with numbers and will run away, chased off the field by a tidalwave of skeletons.

    So I hope I’ve impressed upon you one of the fun and awesome things you can do in dominions. Skellyspam may seem simple, but it’s a fine art of combat and deathly effective when used well.