How ready is Poland in 2026?

The Defence Readiness Index (PWRI) puts Poland's readiness without allies in 2026 at 4.46/10, and 5.07/10 with alliances (NATO) – measured relative to the threat and demands of the era. On the with-allies reading 1683 (5.15) still sits higher, though the gap falls inside the uncertainty bands. Three eras stand high on both readings at once: 1410, 1683 and 2026. What sets today apart inside that trio is not the score but the alliance dimension: a treaty with an Article 5 threshold, permanently stationed troops, prepositioned equipment and joint plans. Allied troops had stood on Polish soil before, from the Tatar auxiliaries of 1655 to the French Military Mission of 1920, but on an incomparably smaller scale and in a looser form. The present-day bottleneck is civil protection.

Is Poland's 2026 readiness unprecedented?

Not without allies. On the reading without them (4.46), 2026 does not stand out from the strongest historical anchors: 1410 (4.75) and 1683 (4.69) score higher, but the index cannot tell any of those differences apart from today. By the same rule 6 of the 9 other columns are indistinguishable from it, 1831 (4.32) among them. Three eras stand high on both readings at once: 1410, 1683 and 2026. What sets today apart inside that trio is the alliance dimension (5.07 with allies, an Article 5 threshold with troops stationed, equipment prepositioned and joint plans), not the composite score.

What does 'readiness ≠ survival' mean?

Higher readiness does not guarantee victory. The standalone readings for the lost war of 1831 (4.32) and the won war of 1920 (4.01) differ by 0.31 of a point, less than the uncertainty band on either of them (1.25 and 1.50). This index cannot tell those two eras apart, and one of them lost while the other won. Alliances, scale and chance decided the outcome, not the numbers alone. The index measures preparation; it does not predict the course or outcome of a war.

Is this a forecast of war?

No. The index does not predict whether war will break out or whether a defence would succeed. It measures defensive readiness at a moment in time, relative to the threat characteristic of the era – conditional analysis, not a probability forecast of aggression.

What sources is the index based on?

On 863 cited sources and 1654 evidence items, organised in four reliability tiers (from T1 official down to T4 other), with mandatory counter-evidence for every dimension and an adversarial red-team review. Historical columns may cite only evidence available at that era's cutoff date (ex ante).

How is readiness measured?

Each era is scored across 10 dimensions – Forces & command, Equipment, Reserves, Logistics, Allies, Strategic shield, Civil protection, Infrastructure, Economy, Intelligence & cyber – on a 0–10 scale relative to the era's state of the art, in two families: with allies and without allies.

Wieki Gotowości The year 2026 against six centuries of history · Defence Readiness Index

How ready is Poland today (2026)? The Defence Readiness Index: a sober, source-grounded reading of Poland's readiness, measured against six centuries (1410–2026 + a 2030 projection), always relative to the threat and demands of each era. One of the three highest readings in the series on both measures, but not a survival forecast.

Thesis: readiness ≠ survival. This index cannot tell the 1831 and 1920 readings apart: less than the uncertainty band on either of them separates the two. One of those years lost and the other won. Outcomes were decided by alliances, scale and chance, not the bare numbers. This is not a war forecast, we measure preparation, not whether war will break out.

Readiness and aftermath – across the ages

Scale 0–10 (relative to the era's threat). "With allies" is the reading with real allied support, "Without allies" the reading by own forces. "Aftermath" described after the fact (firewalled, does not affect the score).

Every score carries an uncertainty band, and the bands of neighbouring eras overlap heavily. A difference of a few tenths of a point therefore does not mean one era scored better than another: only the extremes of the series are distinguishable. The table is ordered chronologically, not by rank.

Year	Era	Threat	With allies	Without allies	Outcome	Aftermath / stakes
1410	Grunwald	Teutonic Order	4.78±1.50	4.75±1.75	mixed	limited – the state survived
1655	The Deluge	Sweden (+ Russia)	2.72±1.25	2.90±1.50	mixed	heavy – demographic collapse
1683	Vienna	Ottoman Empire	5.15±1.25	4.69±1.25	win	limited – the state survived
1792	The Constitution	Russia	3.04±1.50	3.10±1.50	loss	heavy – loss of autonomy
1794	Kościuszko	Russia + Prussia	2.41±1.50	2.56±1.50	loss	catastrophic – loss of statehood – 123 years
1809	Duchy of Warsaw	Austria	3.79±1.25	3.67±1.25	win	limited – the state survived
1831	November Uprising	Russia	4.02±1.25	4.32±1.25	loss	heavy – loss of autonomy – 87 years
1920	Battle of Warsaw	Bolshevik Russia	3.77±1.25	4.01±1.50	win	limited – the state survived
1939	September	Third Reich (± USSR)	3.89±1.50	3.88±1.50	loss	catastrophic – occupation – 6 years
2026	The Present	Russian Federation (+ Belarus)	5.07±1.25	4.46±1.25	open	open – the cost has not fallen due yet
2030	projection	conditional range	3.1–6.3	–	open	conditional stakes

Crises without full-scale war

1956 – Polish October 1956 – threat of Soviet intervention (8th Plenum of the PZPR) (with allies 1.80±1.50) A rough composite (light quota, not to be compared 1:1 with the sound columns of 1410–2026). Deliberately the with-allies reading < the without-allies reading: a perverse 'alliance' (the Warsaw Pact: the aggressor) and Soviet manning of the command lower the adjusted score below the standalone one, rather than raising it. The main calibration takeaway: despite a very low the with-allies reading (~1.8) war did not occur, because the deterrence was decided mainly by the absence of intent for full invasion and Moscow's calculation (plus third parties), not by Polish readiness. This is a warning against naïvely mapping the with-allies reading, hence outcome: a low the with-allies reading does not predetermine war when the adversary has no intent to wage it. Contrast: Hungary a few weeks later, with a comparable will to fight, but Nagy declared exit from the Pact (removing the Soviet 'exit route') and was attacked; Gomułka stayed in the bloc and avoided intervention (an ex post example, outside the dimensional evidence).
1981 – Martial law 1981 – supposedly averted USSR/Warsaw Pact intervention (the Solidarity crisis 1980-81) (with allies 1.40±1.50) A rough composite in a light quota and of weak commensurability; it must not be compared 1:1 with the sound columns. Convention: the with-allies reading computed conditionally on the reference threat (the own hegemon USSR/Pact), which controls the officer corps, doctrine, equipment supply and strategic weapons and keeps troops on PPR territory. On this view the mass and equipment are an asset of the Pact serving the threat, not a defence against it, so an honest composite comes out low and with a wide band. The gap between the without-allies reading (~2.9) and the with-allies reading (~1.4) is itself a finding: standalone, the PPR looked moderately ready, but against its own hegemon defensive readiness collapses to ~1.4, and even that figure is barely commensurable.
2014 – Crimea and Donbas 2014 – risk of the NATO–Russia escalation spilling over onto Poland (with allies 5.50±1) A rough composite (light quota, a directional signal, not to be compared 1:1 with the sound columns of the index). The with-allies reading ~5.5 (band ±1.0) clearly above the without-allies reading ~3.4, consistent with the picture of 'the without-allies reading flat, NATO the differentiator'. Note: the NATO boost is smaller than in the 2026 column, because in 2014 the allied forward presence was thin (before the eFP of 2016/2017); this relative, smaller the with-allies reading−the without-allies reading difference is an honest signal here, not a precise number. Most important: low proximity and the dominance of absence of intent mean that this point should not be read as strong evidence of the effectiveness of the deterrence threshold.
2022 – Full-scale Russian invasion of Ukraine (24 Feb 2022) – Poland a NATO front-line state, no attack on Poland (with allies 5.00±1.50) A rough composite (light quota, a directional signal, not a model): the without-allies reading ~3.8, below the 2026 column ~4.5, since Poland was before the modernisation leap (a smaller army, no delivered Patriot, ~2.2% of GDP); the with-allies reading ~5.0, practically level with the 2026 column (~5.1). The entire the with-allies reading minus the without-allies reading difference (~+1.2) comes from the allied dimensions: d05 (not scored without allies, 7.5 with allies), d06 (+2.5), d01 (+2.0), d04/d09/d10 (+1.0). This shows that the higher reading of this point comes almost entirely from the alliance, but it does not prove that the alliance, rather than national readiness, decided the security of the front-line state. Do not compare 1:1 with the sound columns.
2025 – Drone incident over Poland (9–10 September 2025): a sub-threshold hybrid probe (with allies 6.70±1.60) A rough light-quota composite (a directional signal, not to be compared 1:1 with the index's sound columns). The with-allies reading ~6.7 sits appreciably above the without-allies reading ~5.1, so NATO is the differentiator, especially in d05/d06. The higher the with-allies reading level co-occurs with the incident being kept below the threshold of war, alongside visible solo gaps (low-altitude defence d02, detection d10, civil protection d07). A single case does not establish a correlation.
2026 – A Russian Kh-101 cruise missile near Tarnawa-Kolonia (30 July 2026) – an armed sub-threshold probe that was not stopped (with allies 5.50±1.50) A rough light-quota composite (a directional signal that must not be compared 1:1 with the index's sound columns, nor with the 2026 column, which measures readiness against full-scale war). The with-allies reading ~5.5 against the without-allies reading ~4.2. NATO remains a differentiating factor: the with-allies reading minus the without-allies reading gap here is 1.3 against 1.6 in the September 2025 drone incident (6.7 against 5.1). In this episode the alliance did not prevent the event, responded by air alone, and spoke collectively only after thirteen days, with no new reinforcement measures. The distribution of scores is strongly uneven: the dimensions actually tested by the reference threat (d02 3.5, d06 3.0, d07 2.5, d01 4.0, d10 4.5) sit clearly lower than the background dimensions (d03-d04, d08-d09 around 5-6). The lowest scores on this card concern capabilities directly involved in responding to a single missile. The card does not establish how Poland would fare in a full-scale war.

Rhymes of history

No real alliance leads to catastrophe – Isolation, or guarantees on paper only. All four columns ended in heavy or catastrophic consequences, and readiness in them was low: 2.4 to 4.3 across both readings. So these cases say nothing about what high readiness would have done. (1655, 1794, 1831, 1939)
Real support raises the threshold of aggression – When an ally (or the union’s own mass) is physically present, the threshold of aggression rises. The name is about the threshold, not about survival: survival also depends on scale and chance, and one of these columns is still open. (1683, 1920*, 2026*)
An internal rift completes the outsider’s work – Targowica or the magnates' capitulations: the internal enemy often completes the external one's work. (1655, 1792)
Reforms that come too late do not save the state – Even sound reforms (the 3 May Constitution) cannot save a state when they arrive too late. (1791–1792)
Neglected opportunities leave gaps in the system – The missing Polish connection to the NATO fuel pipeline (CEPS) lowered the infrastructure and military mobility score in the 2026 column. What the index flags as the bottleneck there is something else, civil protection, so this is a gap recorded in the rationale rather than the one the index points at. (2026)
Twice, a collapse was followed by a rebirth – Two cases: 1918 and, in a limited form, after 1945. Neither is in the index, because neither is an era column. Two cases are too few for a rule, and 1794 cost a hundred and twenty three years without a state. (1918, 1945–1989)

* 1920: no physically present combat ally (the advisory French Military Mission, about 600 officers; the Franco-Polish alliance came only in 1921), and the alliance dimension is among the lowest in the series here. * 2026: an open column with no outcome, so it stands there as an instance of the mechanism, not its confirmation.

Methodology

What we measure: The index assesses Poland's defensive readiness at a given moment, relative to the threat characteristic of the era, not in a vacuum. This is not a forecast of whether war will break out, nor whether the defence would succeed.
Scale 0–10 (relative): Assessment relative to the state of the military art of the era: 0–2 defencelessness · 3–4 functions with bottlenecks · 5–6 solid with gaps · 7–8 high readiness · 9–10 aggression becomes irrational. 10 is a model, no one in the series attains it.
What the uncertainty band means: The band on a score is a judgement about the strength of the evidence base, not a confidence interval: no significance level and no probability distribution sit behind it. The anchors are the same across every era column: ±0.5 where the evidence is strong and concordant, ±1.0 in the typical case, ±1.5 to ±2.0 where the base is thin or divergent, with intermediate values where the material calls for them. The band on an era column's composite index is the root mean square of the dimension bands, rounded up to 0.25; crises without war carry a separately estimated band. We do not tell two readings apart while less than the narrower of their two bands separates them; at bands of this width a difference of tenths of a point is not a difference.
Two measures: “With allies” counts real allied support (today: NATO), “without allies” counts own forces only. This is not always the same profile plus one score: an alliance can change what forces, logistics and command are worth. In 7 of the 10 columns it shifts something besides the allied dimension, and for the 2026 column as many as 7, though in some columns it is a difference on a single score. The difference can be negative, because the allied reading adds a dimension that does not exist without allies, allied deterrence, and that one can score below the rest of the defence. What counts is real presence, not guarantees on paper.
Ex-ante discipline: Historical columns are assessed solely on the basis of what decision-makers could have known at the moment of cutoff, without hindsight. The validator enforces this rule at the level of the evidence.
10 dimensions: Forces & command · Equipment · Reserves · Logistics · Allies · Strategic shield · Civil protection · Infrastructure · Economy · Intelligence & cyber. The main result is the profile; the aggregate index is secondary, with an uncertainty band and a sensitivity test on the weights.
How ten scores become one number: The composite index is a weighted mean of the dimension scores, computed with three weight sets at once. The published number is the plain mean of those three results, and the spread between them is the sensitivity test on the weights: for the 2026 column it comes to 0.26 of a point at a reading of 5.07 with allies, and 0.15 of a point at 4.46 without. In the standalone reading the allied dimension is not scored, and the weights of the remaining nine are renormalised. The full sets, each summing to one hundred per cent: equal weights: Forces & command 10%, Equipment 10%, Reserves 10%, Logistics 10%, Allies 10%, Strategic shield 10%, Civil protection 10%, Infrastructure 10%, Economy 10%, Intelligence & cyber 10%; combat-capability emphasis: Forces & command 15%, Equipment 15%, Reserves 10%, Logistics 15%, Allies 10%, Strategic shield 8%, Civil protection 5%, Infrastructure 10%, Economy 7%, Intelligence & cyber 5%; resilience emphasis: Forces & command 7%, Equipment 7%, Reserves 10%, Logistics 12%, Allies 8%, Strategic shield 7%, Civil protection 17%, Infrastructure 12%, Economy 10%, Intelligence & cyber 10%.
Cost and consequences – after the fact: The human cost and long-term consequences are described ex post and firewalled: they never feed, sort or weight the readiness assessment. This is a separate layer of memory, not a measure.
Sources and counter-evidence: Every piece of evidence has a citation, a date and a tier (from T1, official, down to T4, the rest). For each dimension countering evidence is mandatory: a dossier without counter-evidence does not pass the check. Links are checked; for dead ones we give an archive. The base is predominantly Polish-language: 652 of 863 registry entries, against 192 in English. The remaining 19 are German (10), French (4), Russian (2), Swedish (2) and Latin (1). A reader without Polish can therefore check the figure and the path to the source, but usually not the source itself.
Adversarial review (red team): Three independent adversarial passes, run by a model rather than by people, were told to overturn the scores instead of confirming them: they hunted for badly placed anchors, missed counter-evidence, breaches of ex-ante discipline and bands set too narrow. The record sits on 27 of the 110 dimensions of the whole scoring corpus, spread across 11 columns, and covers 10 of them; the only column with no entry is 1792. There are 35 entries. Twenty-one of those 27 dimensions can be opened on this page; the difference is what the page does not show as a separate column: 1939B, the scenario variant of 1939. Thirty-two record a recalibration of a score, a band or a cap, five record a challenge considered and rejected with reasons, and two belong to both groups at once: the central value was defended and the band widened instead. The text of the challenges on the dimensions this page opens can be read in the components drawer. Beyond those, the record sits in the reviewer's working shorthand next to each dimension in the scoring layer, with a summary of some columns in the review log red-team-log.md; the repository is private, so that part cannot be checked without access to it.
Crises without war are scored differently: Crises without war carry a lighter evidence quota than the era columns, and a rough composite instead of the full scoring procedure. They are a directional signal, not a measurement on the same scale. A crisis reading (2025: 6.70 with allies) can sit above every era column (the peak of the series is 1683: 5.15), and that does not mean readiness was greater that year. A crisis measures resilience against one sub-threshold event, an era column measures readiness against full-scale war. The two sets of numbers should not be placed side by side.
Limitations: The index measures the development of the state apparatus relative to the scored dimensions, not survival against an overwhelming advantage of scale (that lives in the unscored threat context). This is why the index cannot tell the reading of the lost year 1831 apart from that of the victorious 1920. Nor do the ten dimensions score separately several functions on which a state's ability to operate under strain depends: people's psychological resilience, the continuity of basic services (water, food, medical care, payments) and dependence on space services such as satellite navigation and communications. They appear in the evidence of broader dimensions, but none has a score of its own, and in the 2026 column health care currently has no evidence admitted to scoring. A high dimension score therefore does not settle whether these functions would hold up in a war.

Sources

Based on 863 sources and 1654 evidence items (tiers: T1 107 official · T2 269 research · T3 416 journalism · T4 71), with mandatory counter-evidence. Links checked: 846/863 live. The base is predominantly Polish-language: 652 of 863 registry entries, against 192 in English. The repository is private, so the address is given for the record: https://github.com/kgluszczyk/worrier.

About the project

This index arose as a continuation of the author's private attempt to anchor his anxiety about the impact of war on himself and his family in Poland, to give it measure, context and history, instead of the fear alone. Co-created with an advanced AI (Fable 5): source research, dimensional assessment and the texts emerged in a human–model dialogue, under human oversight and under hard rules of honesty (ex-ante discipline, human-cost firewall, mandatory counter-evidence, red-team review).

State of knowledge: the “today” column (2026) and the projections are frozen as of 29 August 2026. Historical columns are assessed ex ante, solely on what was knowable on that era's cutoff date.

For all who had to stand ready.

Wersja polska · Terms, copyright and privacy

This page is interactive. Enable JavaScript for the immersive version with charts and era galleries.

Wieki GotowościDefence Readiness Index