Forecast

MCARI Forecast Bulletin · No. 1 · October 2026

Forecast for January–March 2026

MCARI data, release v1.1 · data to 31 December 2025

Among states and regions with at least 50 events last quarter, none is forecast to rise in January–March 2026; the largest fall is in Tanintharyi (−33%).

4,214

events forecast nationwide (sum of the 18 most likely values)

5,639

No change (last quarter): the same if nothing changes from last quarter

Pending

the last forecast is scored when all three months of its quarter are recorded

Limits on forecasting. MCARI forecasts at state and region level only. It publishes no forecast for a township or any smaller place, so that nothing it publishes could help any party choose where to strike. It does not forecast what any named person or group will do. Its forecasts are published for humanitarian and civilian-protection purposes only.

MCARI records conflict events reported in the news. Where reporting is thin, counts, and so forecasts, can be too low.

Where conflict is forecast

Most likely number of events

<40
40–99
100–179
180–299
300+

Forecast by state and region

Events and Civilian deaths, with the 80% range, the no-change baseline and the last eight quarters

State / region	Events, Most likely	80% range	No change (last quarter)	Against baseline	Chance of a rise	Civilian deaths (80% range)
Sagaing	875	438–1,530	1,209	▼ 28%	23%	161 (41–403)
Magway	562	288–1,000	767	▼ 27%	27%	81 (22–200)
Mandalay	474	248–821	648	▼ 27%	26%	117 (29–284)
Kayin	278	138–490	391	▼ 29%	23%	29 (8–76)
Rakhine	231	118–395	303	▼ 24%	27%	87 (23–211)
Kachin	225	111–398	297	▼ 24%	28%	34 (8–86)
Bago-West	216	106–376	280	▼ 23%	28%	27 (6–67)
Tanintharyi	212	108–379	318	▼ 33%	19%	16 (4–42)
Yangon	175	90–308	237	▼ 26%	25%	6 (1–17)
Shan-North	159	81–281	204	▼ 22%	29%	60 (15–153)
Chin	145	70–255	175	▼ 17%	33%	20 (5–55)
Bago-East	145	72–253	177	▼ 18%	32%	23 (6–57)
Mon	130	67–234	166	▼ 22%	31%	18 (4–47)
Shan-South	103	51–176	123	▼ 16%	35%	15 (3–37)
Kayah	97	47–172	125	▼ 22%	30%	17 (4–43)
Ayeyarwady	96	47–169	118	▼ 19%	33%	28 (7–71)
Naypyitaw	84	41–147	95	▼ 12%	39%	2 (0–6)
Shan-East	7	3–15	6	▲ 17%	58%	2 (0–6)

Bar: the 80% range on a scale running to the largest value in the column; dark tick: the most likely value. Sparkline: recorded events, 2024 Q1 – 2025 Q4. Chance of a rise: the probability that the quarter has more events than last quarter (the No change column).

How did the last forecast do?

The January–March 2026 forecast, made from data to 31 December 2025 (MCARI data, release v1.1)

Pending — scored when January–March 2026 are recorded

How accurate have we been?

The quarterly model behind this forecast, tested on every quarter 2022–2025, one quarter ahead, against “no change from last quarter”.

	Skill against no change	Inside the 80% range	Mean absolute error Model / No change	Root mean square error Model / No change
Events	+0.20	84%	65.5 / 63.5	117.7 / 106.0
Civilian deaths	+0.16	86%	25.8 / 23.1	52.2 / 45.0

Skill: how much closer the whole forecast, range included, comes to what was recorded than “no change” does (0 = no better, 1 = perfect). Errors: the most likely value against what was recorded, per state or region per quarter.

Share of the 18 inside the 80% range, every quarter scored. Hollow: Backtest. Filled: Published bulletins. One point is added each quarter. Dashed line: 80% intended.

Spotlight

Events with recorded displacement

Recorded events by month, January 2024 – December 2025; past events only, no forecast

Peak month: October 2025 (131). Busiest state or region, October–December 2025: Sagaing (78).

Download PDF · Requires an account.

What changed in this bulletin. First issue: the method is as published on page 3.

Cite as MCARI (2026). MCARI Forecast Bulletin No. 1: January–March 2026. Myanmar Conflict Analysis Research Institute.