Will ACLED report any conflict fatalities by India against Pakistan or vice versa between May 13 and the end of May?
Jeremy’s Multi-AI Oracle predicts 35%
Obtaining outside data.
Checking internet sources, using keywords: ACLED conflict fatalities India Pakistan May 2023
Results from around the web:
Okay, let’s tackle this query. The user wants current web results related to ACLED conflict fatalities between India and Pakistan in May 2023, along with data tables and sources. First, I need to check the provided search results for relevant information.
Looking at the sources,
[1] mentions over 147,000 fatalities worldwide up to July 2023, but it’s a global figure.
[2] is more specific: it talks about a 25-fold increase in political violence in India’s Manipur in May 2023 due to ethnic clashes. That’s in India, so relevant. Also, in Jammu & Kashmir, clashes doubled with 10 events, including an IED attack killing 5 soldiers. PAFF, linked to Pakistan’s Jaish-e-Mohammad, was involved. So fatalities in India’s J&K are part of the India-Pakistan conflict.
(Carolyn‘s note: India’s Manipur is at the far eastern borderlands of India, so not relevant.)
Source [3] discusses Pakistan’s violent 2024, which is outside the May 2023 timeframe.
[4] mentions a 2025 attack in Kashmir, again not relevant.
(Carolyn’s note: false, Kashmir is relevant, as control over Kashmir is de facto divided among India, Pakistan, and China.)
[5] is a PDF on Pakistan’s 2023 incidents. The PDF data shows fatalities in various provinces, including Azad Kashmir (5 fatalities) and others, but Khyber Pakhtunkhwa had 1,356 fatalities. However, this is Pakistan’s internal data, not cross-border with India. But since the query is about India and Pakistan, I should include both countries’ data even if internal, as cross-border influences might be implied.
Now, compiling the data: For India in May 2023, Manipur had a surge, but ACLED’s [2] notes 25-fold increase in political violence, but exact fatalities? The J&K part had 10 events, with at least 5 fatalities from the IED attack. The PDF [5] for Pakistan in 2023 lists fatalities per province, but May-specific data isn’t clear. The user asked for May 2023, so I need to check if the sources have monthly data. [2] is May 2023 for India, [5] is annual 2023 for Pakistan. So for India, May data is available, for Pakistan, annual. But the user might want the closest available.
Tables: For India, from [2], May 2023 in J&K: 10 events, 5 fatalities. Manipur’s 25-fold increase but no exact numbers. For Pakistan, using [5], annual 2023 data by province. Since the user wants time series, but the sources don’t provide monthly breakdowns except for May 2023 in India. So the time series might be limited. The change from previous period (April 2023) for J&K is doubled events, but fatalities? The IED attack had 5, but total for May isn’t specified beyond that incident. Maybe the 5 are the total for J&K in May.
So the summary would highlight the key incidents in India (Manipur and J&K) and Pakistan’s overall 2023 data, noting that May-specific data for Pakistan isn’t available. Include tables for India’s May figures and Pakistan’s annual, and list the URLs used.
Political violence in May 2023 saw significant developments in India and Pakistan, with distinct regional patterns. In India, inter-ethnic clashes and cross-border militancy drove fatalities, while Pakistan experienced persistent militancy-related violence.
### India: May 2023 Conflict Fatalities
**Manipur** (Carolyn’s note: This province is at the far eastern boundary of India, whereas Pakistan is at the far western boundary. Therefore, Manipur fatalities are irrelevant.)
– Inter-ethnic clashes caused a **25-fold increase in political violence** compared to April 2023[2].
– Exact fatality counts aren’t specified in available data, but ACLED recorded **over 139,000 global political violence events** in this period[1].
**Jammu & Kashmir**
– Clashes between security forces and militants doubled, with **10 events** recorded[2].
– A 5 May IED attack by the Pakistan-linked People’s Anti-Fascist Front (PAFF) killed **5 soldiers** in Rajouri district[2].
—
### Pakistan: 2023 Conflict Overview (Annual Data)
*Limited monthly Pakistan data available for May 2023[2][5].*
—
**Sources Used**
1. https://acleddata.com/conflict-index/index-july-2023/
2. https://acleddata.com/2023/06/09/regional-overview-asia-pacific-may-2023/
5. https://www.ecoi.net/en/file/local/2107103/2023yPakistan_en.pdf
OpenAI gives a base rate of 0.1
The question difficulty is rated 7 (0 to 10)
Historical weighted factors include:
India-Pakistan conflict history, 0.3
Recent increase in regional violence, 0.4
Cross-border attack frequency, 0.3
A Bayesian calculation could be performed as follows:
Using prior probability from historical conflict data (0.1), adjusting with likelihood of recent May 2023 violence (0.5) and absence of major border conflict news (0.6): P(A | B) = (0.1 * 0.5) / ((0.1 * 0.5) + (0.9 * 0.4)) = 0.122.
Bayesian base rate: 0.122
Sufficient news to provide a good forecast? 0 (0 or 1)
News is relevant, topical and unbiased? 1 (0 or 1)
Question classification: scenario_based_forecast
Expected time (days) until question close: 30
The following were considered in order to produce this base rate:
The base rate is derived from historical conflict occurrences between India and Pakistan, adjusted for a noted increase in regional violence, such as the drastic rise in Manipur in May 2023.
Ideally, the news feed would contain the following sorts of information for a better forecast:
Detailed monthly fatality data for both countries in May 2023, plus explicit identification of cross-border incidents would improve the forecast.
Some potential divergent considerations that might affect the base rate:
Potential underreporting or lack of detailed monthly fatalities for Pakistan in May 2023 could alter the perceived level of cross-border conflict.
The following chain of events are necessary for the question to resolve positively:
– Intensified political violence in Manipur, India with specific fatality counts known. (Unlikely)
– Increased conflict events and fatalities in Jammu & Kashmir, India compared to previous months. (Moderately likely) – Significant cross-border incidents between India and Pakistan leading to fatalities. (Unlikely)
Querying Claude (AI predicts: 0.35 – confidence: 6)
Querying Mistral (AI predicts: 0.15 – confidence: 5)
Querying OpenAI (AI predicts: 0.65 – confidence: 6)
Question Type: Binary
Median from LLMs: 0.35
Base rate: 0.1 (from OpenAI)
SD: 0.21
MAPD: 0.33333333333333
Confidence: 6
Conf Mode: Low
Mellers: 0.29
Reverse Mellers: 0.4
Theory of Mind: 0.5 (What did the LLMs think other LLMs predicted?)
Beta Distribution: 0.08
Close Type: B (B = cautious # closer to 50%; A/C = closer to extremes)
# LLM responses: 3
Definitions of the above statistical measures are at the bottom of this page.
Model value: 35%
The various AI analyses converge on the notion that predicting conflict and fatalities between India and Pakistan is fraught with uncertainty, hinging especially on historical patterns, regional tensions, and current geopolitical dynamics. There is agreement that historical data indicates a generally low base rate of conflict fatalities, possibly complicated by recent increases in regional violence, especially in Jammu & Kashmir, and domestic unrest such as that in Manipur. The short timeframe for this prediction limits certainty, and there is an inherent unpredictability in geopolitical tensions that could lead to unexpected escalations. Furthermore, potential underreporting of conflict incidents by sources like ACLED adds to the uncertainty, and the persistent underlying tensions, particularly in regions like Kashmir, imply that while significant cross-border conflicts are currently unlikely, sudden changes in political or military actions could alter the situation rapidly.
Runtime: 93 seconds.
Links to all past forecasts —>
Sources: graphics from https://www.scmp.com/ and https://x.com/Sturgeons_Law/status/1920341134435344510 as presented on Slow Boring https://www.slowboring.com/p/were-sleeping-on-the-most-dangerous
Guide to Jeremy’s Bots Reports
Median: We collect the output values from the LLMs in an array. We then take the median value. This is between 0 and 1 (i.e. 0.15 = 15%).
Base rate: Currently we take OpenAI’s value as the base rate. We are working on a better value using a different prompt that takes into account process. The base rate helps to determine if the median makes sense.
SD: The standard deviation between the median and the base rate.
Confidence: We query each of the LLMs on how confident they are of their predictions (between 0 and 10) and take a median. Because the LLMs tend to be overconfident, we take anything lower than 6 as being low confidence. This factors into the overall model.
Conf Mode: Based on the confidence value. >=9 is high confidence. Below 6 is low confidence (this is also triggered by an exceptionally high SD).
Mellers: This refers to Barbara Mellers, specifically a paper she wrote that includes a formula for moving values towards an extreme (i.e. 0 or 1).
Reverse Mellers: This uses the formula from above, but with a sub-1 coefficient to move the values closer to 50%.
Theory of Mind: We ask the LLMs what they think other LLMs would predict. We hope that this makes them consider the questions more deeply.
Beta Distribution: Currently unused, but possibly of interest. This is based on the median, the base rate and the SD.
Close Type: We noticed that the appropriate base case for some questions is closer to the extremes, while others are closer to 50%. When we have a low confidence value, this helps us to determine whether to extremize or de-extremize the value. ‘A’ implies closer to zero. ‘B’ implies closer to 50%. ‘C’ implies closer to 100%.
# LLM Responses: The count of LLMs that responded. We query 5 currently, but they have a relatively high API failure rate.
Model value: For normal confidence mode, this is the median. For high confidence, this is the Mellers value. For low confidence, this is either the Mellers or Reverse Mellers (depending on the Close Type).
