tonio

Today, gunfire erupted along the border resulting in one civilian death, confirming (1) the hair-trigger environment and (2) that combat can occur despite international monitoring – but both sides stopped after only 10 minutes with remarkably low lethality.

Updating in light of the landmine blast that injured four Thai soldiers – which puts this in a critical period of potential escalation. Beyond this crisis, my 14% also incoporates additional possible landmine incidents over the next 62 days. Factors favoring restraint are (1) massive economic interdependence, (2) active ASEAN monitoring, and (3) US officials direct engagement. So far we have seen Thailand respond is a way that is relatively measured – suspension, then demands, then deployment rather than immediate attack – which could sugguest signaling deterrence rather than an imminent offensive. Going to be watching this closely over the next 72 hours... 

There are several low-cost pathways for implementation (e.g., entity list additions) that provide high aggregate probability even without a comprehensive new rule. Besides those, the GAIN AI Act passed the Senate in October 2025 – and could include mandates for BIS. Finally, BIS explicitly promised replacement rules, which indicates continued regulatory intent rather than abandonment of controls.

Base Rate

From my quick research I identified four major publicly signed contracts over €1B with non-European suppliers in 2024-2025 (1/24, 8/24, 11/24, 7/25). While this is only a quick search, from it I assume a rate of roughly 2-3 major contracts/year across European NATO members. For the forecasting period, a continuation of that pace would suggest an expectation of 1.4-2.1 contracts – supporting a baseline probability of 50-65% for at least two contracts. However, I adjust that for (1) potential front-loading as countries may have completed urgent "first wave" procurement in 2022-2024, (2) risk of current agreements slipping outside the forecasting window, and (3) intensifying European procurement preference policies. Altogether, I assume a base rate of 52-58%. 

Why I Think I'm Right

There are several avenues for the question to revoles 'Yes'. First, it appears Germany's planned purchase of 15 additional F-35s has strong momentum and the pace of Poland's procurement sugguest more deals could be on the way. Further, Greece appears interested in an Israeli air defense system, and we could see other large NATO members sign deals. When combining structural demand drivers (NATO's 5% GDP commitment, capability gaps in F-35/air defense with no European alternatives, ongoing threat environment) and multiple paths to resolution supports a modest tilt toward 'Yes'.

RSF forces are currently ~200km from Khartoum and must capture multiple heavily-defended intermediate cities (al-Obeid, Bara) before reaching the capital. Further, SAF is now an entrenched, prepared defender with fortified positions and 20,000+ troops. Looking at OSINT, it appears the RSF is consolidating its el-Fasher victory rather than preparing a major new offensive. Finally, the resolution criteria requires "sustained control" through March 31, 2026 – making the bar even higher for this to resolve Yes.

Base Rate

Looking at previous recognitions, they happen in waves during major diplomatic events – 6 countries in 1988, 2 in 2010, and most recently 5 in September 2025 coordinated around the UNGA. The remaining holdouts represent the hardest cases: either ideologically committed to Israel (US, Germany), dependent on US security (Japan, SK), or led by right-wing governments (US, Italy). Outside of waves, the base rate for G20 recognition in any given 3-month period is historically <1% per country/quarter.

However, Italy's parliamentary motion conditionally recognizing Palestine upon "all hostages" being released creates a time-sensitive catalyst absent from the historical base rate. With Phase 1 of the Gaza ceasefire complete and Phase 2 ongoing, I assume a 25-35% probability of Italy's conditions being met in the forecasting window. However, even with Italy, two other holdouts would need to recognize Palestine to exceed 16. 

Why I Think I'm Right

The forecasting timeframe is relatively short (88-days) – and the remaining holdouts chose not to recognize during the September 2025 wave. Italy is the most likely to recognize Palestine, though that is subject to hostage remains being returned and Meloni choosing to follow through. Even if Italy recognizes, cascade effects are likely limited – Japan's new PM Takaichi and SK's pro-alliance NSA make follow-on recognition within weeks unlikely without a coordination mechanism. Reaching 17 total countries requires three simultaneous new recognitions in 88 days, which demands either a highly successful cascade or an extreme external event, both of which are tail scenarios properly allocated at 1%.

Base Rate

I derive a base rate by considering (1) historical reference classes and (2) mechanisms for escalation. 

• Historically, ceasefires between Cambodia-Thailand appear to be quite fragile (2008-2011: 8 ceasefires, 2025: 3 ceasefires) – but that represents the rate of violations not mass casualty events. In the last 30+ years, years with disputes exceeding 20 deaths were only 2011 and this year, yielding an annual base rate of 5.5%. 

• Mechanistically, one could assume ~4 landmine incidents in the 73-day window (1.8 incidents/month x 2.37 months) each with an assumed 2% probability of triggering escalation to 20+ deaths. This gives an ~8% chance of a landmine-trigged escalation. Adding in smaller probability scenarios (e.g., deliberate escalation) may yield a base rate of ~10-12%. Given the uncertainty I have around these parameters and the monitoring infrastructure, I discount that base rate by 50% – giving a 6% base rate. 

Weighing these, I assume a base rate of 5-6%.

Why I Think I'm Right

As of forecasting, the Kuala Lumpur Peace Accord is less than two weeks old – but includes genuinely unprecedented elements (e.g., ASEAN observers, heavy weapons withdrawal). Under the current monitoring regime, there have been zero incidents which have escalated, but my 6% forecast balances this structural change against real but manageable tail risks. 

I thought about a few different reference classes: 

First, annual Q4 volumes. Q4 2022-2024 averaged ~12680 MCM. Given the Q3 volumes (8167 MCM), Q4 requires 10833 MCM to hit 19000 MCM. 

Second, YoY volumes. Q1-Q3 2025 showed accelerating YoY declines (-28%, -28%, and -39%), and reaching 19,000 MCM requires Q4 to decline only -23% YoY from Q4 2024. 

Third, the October 2025 data. October saw approximately ~2,500 MCM (LNG + TurkStream), requiring November-December to average 4,164 MCM/month.

Given that – and infrastructure constraints – I put the probability of hitting threshold at 3-5%.