Country Scoring

We started from the same decision here as we did with weapons. Instead of running a formula over a country's stored attributes, like troop counts, tank counts, or budget figures, we ask the AI what it genuinely knows about that country. The reason is the same too. Our country data is still incomplete, and in places wrong, for most countries. A formula fills that gap without noticing it is there, and hands you a number that looks reliable but is hollow underneath. We cut that bridge on purpose. Scoring never looks at the attributes stored in our country database. It comes entirely from the AI's own knowledge.

We score across ten separate headings. Land power, air power, naval power, leadership, strategic deterrence, industry, intelligence and command systems, logistics, economy, geostrategic position. Each one is judged on its own, based on what is genuinely known about that country.

We hold to the same principle here. Evidence, not reputation. A country's reputation is not always an accurate reflection of its real capability. Saudi Arabia buying the world's most expensive weapons does not make it a strong military. The long campaign in Yemen exposed real command and planning weaknesses. Azerbaijan, on the other hand, is widely seen as an authoritarian, non-NATO country that rarely makes Western headlines. But in two wars against Armenia it used drone and precision-strike integration better than many far more recognized militaries. When we score, we look at what actually happened in the field, not at reputation.

But reducing a country's power to one score carries a difficulty that was never this sharp with weapons. Defensive strength and offensive strength are not the same thing. A country can be formidable at holding its own ground and still be incapable of projecting power beyond its borders. Or the reverse, a military that can strike far away while struggling to hold its own frontier. We try to capture both by spreading them across separate components, but everything still collapses into one number in the end. That collapsing is itself a simplification. We are not hiding that.

We also want to say this plainly. Our score is not a prediction of who would win if two countries went to war. When you see Russia at sixty seven and Sweden at forty, that gap does not translate into some equivalent ratio on a battlefield. What decides the outcome of a war is far more than a power score. Who is attacking, who is defending. What the terrain looks like. Which allies show up. Which side's population is actually willing to fight. None of that fits into a single power number. Our score is a summary of a country's overall military capacity, not a simulation of what happens when two countries collide.

Being visible in the media does not make a country stronger either. However often a country turns up in the news, that does not add a single gram of strength to its military. Sometimes it is the opposite. A country constantly in the headlines may be making noise precisely to cover a weakness. A quiet country may simply be strong enough that it feels no need to prove anything to anyone. When we score, we are hunting for real capability, not visibility.

Raw numbers are evidence too, not perception, and we try not to forget that. A country's troop count, tank count, or aircraft count is a real difference in scale, and we cannot look past that when scoring. North Korea has over a million soldiers and a massive concentration of artillery within range of Seoul. That is a real scale, not a reputation. We did not fully capture this in our earliest attempts. In one pass, North Korea's land power came out equal to Sweden's, a country whose active army does not reach seven thousand people. Once we noticed, we went back and fixed it. Raw scale is never a detail we can afford to overlook.

We also made the opposite mistake, and we are not going to hide that one either. In our next attempt, the effort to correct for reputation swung too far the other way. We pulled Russia's air, naval, industry, and economy scores down so hard that its overall score fell below Sweden's, landing almost level with Germany's. That was not a defensible outcome. Russia is a nuclear superpower. Its economy and industrial base are several times larger than either Sweden's or Germany's. It has real weaknesses, and we are not looking past them, but those weaknesses are not large enough to put it on par with a small European country. We thought about what was behind this mistake too. AI models are trained largely on English-language, Western-media-heavy sources, and that can quietly build in a benefit of the doubt toward NATO and Western countries, and an equally quiet negative framing toward countries like Russia and China. Once we noticed this, we added an explicit check for every country. We asked the model to consider whether it would score the same way if that country's political standing in Western media were reversed. Running into three mistakes like this back to back shows the system is fragile. It also shows we can find our mistakes and fix them. Both are true.

A country's strength is not just what sits in its arsenal. Internal dynamics matter just as much as the inventory itself. A stable country does not always fight better. Sometimes a population with almost nothing left to lose fights with a resolve and intensity that a well-equipped, well-funded army cannot match. History is full of examples. We try to fold factors like this into components such as leadership, but we know this is not an exact science. No formula can fully predict when and why a society finds the will to fight.

We also want to keep it visible how the ten components feed into the overall score. Each component starts from a base weight. Land power fourteen percent, air power fourteen percent, naval power nine percent, strategic deterrence thirteen percent, leadership seven percent, industry ten percent, intelligence and command systems twelve percent, logistics nine percent, economy seven percent, geostrategic position five percent. We did not take these numbers from any outside source. They come from our own reasoning. We deliberately kept logistics and command-and-control systems above naval power and economy, because in modern conflicts these are the domains that actually decided outcomes, not the more impressive-looking hardware totals.

But these base weights do not stay fixed for every country. They shift based on four real, structural facts. If a country is landlocked, its naval weight drops sharply, because no amount of theoretical naval strength makes it a coastal power. If a country holds a complete nuclear triad, its strategic weight rises. If a country's defense industry is fully independent, its industry weight rises. If a country faces active border disputes and multiple nearby rivals, its geostrategic weight rises. We do not keep these weights in a black box. For every country, which weights were used and why is documented openly in our database.

A country's power is not a frozen thing. An election result, a commander's retirement, a scandal, a new countermeasure fielded against a drone, an aircraft crash, a decline in population growth. Any of these can affect real military strength, and through it, the score. When something significant like this happens, we try to go back and reassess that country. This is not an instant, automatic system. But we believe scores need to move over time as the real world moves.

Our current evaluation method works the way we have described here, but we are not stopping at this. By the end of summer, a new layer goes live that tests country scores against real conflict scenarios. It will run integrated with our weapon scoring system, comparing inventories from the same category and adding a fresh verification layer on top of what is here now. The country scores you see today came out of the first full pass of this system, which makes them a draft. Once the new layer is ready, we will update both this text and the scores together.

So read the scores here as the best judgment we could produce with what we know, open to challenge, not as settled fact. A country score you see in 2026 can change substantially by 2027, because new evidence surfaces, because a country performs unexpectedly in a real conflict, or simply because we made a better judgment call the second time around. If a score looks wrong to you, write to us. The readers who catch the most mistakes are the best company we have on this road.