ResBench

Only detailed assessors' reports are shared with applicants. So, there was no training of that kind.
The app provides feedback from the perspective of an imagined general assessor (the researchers who make up panels)—a well-read non-expert with deep knowledge of ARC funding schemes.

Great to hear we're reaching the right people and you're finding the posts helpful!

This app might be something for you. There's a giveaway in the thread.
bsky.app/profile/resb...

🎁 𝗟𝗮𝘂𝗻𝗰𝗵 𝗴𝗶𝘃𝗲𝗮𝘄𝗮𝘆
ResBench is has three free vouchers, each providing one hour of access.

The first three commenters who complete all steps will be contacted directly with a voucher code.

To be eligible:
• Like the app's launch post
• Repost it
• Follow ResBench
• Comment below.

The app is proof of concept and a finished product, tested by researchers. It is intended to complement—not replace—expert peer review within relevant research fields, while helping to reduce the time and cost burden associated with preparing compliant and competitive ARC grant applications.

Engagement with the virtual assessor provides general feedback on project quality and innovation and anticipated benefits. Additionally, it generates draft content that adheres to ARC instructions:

• Application title
• Application summary
• National Interest Test (NIT) statement.

Virtual ARC General Assessor+ by ResBench Virtual ARC General Assessor+ is an AI-powered virtual general assessor and grant writer for Australian Research Council (ARC) grant applications. The virtual assessor guides the user through Socratic...

Introducing the 𝗩𝗶𝗿𝘁𝘂𝗮𝗹 𝗔𝗥𝗖 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗔𝘀𝘀𝗲𝘀𝘀𝗼𝗿+.

virtual-arc-general-assessor.resbench.com.au

The app guides users through AI-powered Socratic questioning to refine a research project idea for enhanced clarity and alignment with common elements of key assessment criteria.

Data-driven benchmarking underpins @resbench.bsky.social’s work, setting it apart from many research grant consultancies. While we value qualitative insights from institutional knowledge and anecdotes, we also use programmatic collection and computational analysis of publicly available data.

Excluding PIs, successful projects had mostly teams of one—one!—to three CIs, with a sharp decline beyond that.
DP26 (CIs ONLY):
Team of 1 → 104 grants (19.8%)
2 → 174 (33.1%)
3 → 137 (26.0%)
4 → 70 (13.3%)
5 → 27 (5.1%)
6 → 10 (1.9%)
7 → 3 (0.6%)
9 → 1 (0.2%)
Total: 526
Average: 2.6
Median: 2.0

DP26 (ALL CIs & PIs):
Team of 1 → 42 grants (8.0%)
2 → 136 grants (25.9%)
3 → 148 (28.1%)
4 → 88 (16.7%)
5 → 58 (11.0%)
6 → 30 (5.7%)
7 → 12 (2.3%)
8 → 8 (1.5%)
9 → 2 (0.4%)
10 → 1 grant (0.2%)
11 → 1 grant (0.2%)
Total: 526
Average: 3.36
Median: 3.0

Are smaller teams genuinely more competitive, or do these patterns simply reflect received wisdom and institutional norms? During pre-award checks, did your institution comment on team sizes, or did you advise colleagues, and if so, on what basis?

In DP26, successful teams averaged 3.36 investigators including CIs and PIs; counting CIs only, the average was 2.6. DP25 averages were almost identical. Data on DP26 team size distribution show common practices, not success or competitiveness conditional on size.

ARC Discovery Projects 2027 (DP27) expressions of interest are already under assessment, so investigator teams cannot change. With no further round expected, now is a useful moment to reflect on whether past team composition could have been more strategically planned.

New GO published by Australian Research Council: GO8166 - Discovery Early Career Researcher Award for funding commencing in 2027. View the GO at https://www.grants.gov.au/Go/Show?GoUuid=2de2aa29-dbb7-4a53-ab5e-f0eb8aaac69e

ResBench ResBench creates innovative apps, tools, and resources for research grant application development.

If you want to support ResBench's work, or just say thanks, you can pay for our next coffee at buymeacoffee.com/resbench.

Follow @grantconnectbot.bsky.social on Bluesky to keep track of newly published research-related grant opportunities (GOs) from grants.gov.au. Made by @resbench.bsky.social.

@resbench.bsky.social is now on LinkedIn! Find us at www.linkedin.com/company/resb....

Next steps?
1. Consider your FoR code options.
- Use FoR Finder (resbench.github.io/FoR-Finder/).
2. Check your FoR codes’ (up to three) distribution to panels.
- Use FoR-to-ARC-Discipline-Panel Lookup (resbench.github.io/FoR-to-ARC-D...).
3. Evaluate strategic risks and benefits.

Comments welcome!

Strategic implications: bridging disciplines is often encouraged in research, even in project descriptions for submission to the ARC, but the DE26 fellows tended to demonstrate a clear core discipline in FoR code nominations.

While most aren’t crossing panels, their FoR code nominations aren’t "narrow" either. 13.5% (27/200) nominated only one FoR code, whereas 33.5% (67/200) nominated two and 53% (106/200) nominated three.

Here is the breakdown:
- 82.5% (165/200) of fellows had their FoR code nominations distributed one panel.
- Only 16.5% (33/200) crossed two panels.
- Just 1% (2/200) nominated codes distributed to three panels.

@resbench.bsky.social has been crunching the numbers on ARC Discovery Early Career Researcher Award (DECRA) 2026 (DE26) outcomes.
Attention DE27 applicants: analysis of 200 DECRA fellows’ fields of research (FoR) code nominations shows a clear trend—disciplinary focus.

Here’s the breakdown:
- 82.5% (165/200) of fellows had their FoR code nominations distributed one panel.
- Only 16.5% (33/200) crossed two panels.
- Just 1% (2/200) nominated codes distributed to three panels.

While the ARC funds interdisciplinary research, the data suggest that nominating FoR codes distributed to a single discipline panel is most common among DE26 fellows.

See all DE26 outcomes here: tinyurl.com/56j83hja