compositional.mle
An MLE package rebuilt around composable solvers, then renamed to match.
Comparison · Analytics
hubEvals vs mcptools
A side-by-side editorial comparison of hubEvals and mcptools — release velocity, themes, recent moves, and the top alternatives to consider.
Shared themes:r-package
hubEvals vs mcptools: at a glance
What is hubEvals?
Forecast-hub scoring that learned to handle joint, sample-based predictions.
hubEvals scores model output from collaborative forecasting hubs, wrapping scoringutils and translating hubverse formats into forecast objects it can evaluate. The package has moved quickly from a thin translation layer to something that handles every output type the hubverse defines — quantile, mean, median, nominal and ordinal pmf, and samples. The most recent releases are almost entirely about the failure modes of relative skill scoring rather than about new metrics.
What is mcptools?
R became a deployable MCP server, not just a local one.
mcptools lets R act as an MCP server exposing tools to LLM clients, and lets those clients reach into running R sessions. It reached 1.0.0 in mid-2026 after a rename from acquaint and a reversal of its dependency relationship with btw. The trajectory of its releases tracks the MCP specification closely — transport options, protocol version negotiation, and content types have each arrived as the spec settled them.
hubEvals vs mcptools: editorial side-by-side
hubEvals
2.5ANALYTICS
Forecast-hub scoring that learned to handle joint, sample-based predictions.
◆ Current state
hubEvals scores model output from collaborative forecasting hubs, wrapping scoringutils and translating hubverse formats into forecast objects it can evaluate. The package has moved quickly from a thin translation layer to something that handles every output type the hubverse defines — quantile, mean, median, nominal and ordinal pmf, and samples. The most recent releases are almost entirely about the failure modes of relative skill scoring rather than about new metrics.
◆ Where it's heading
Two threads dominate. The first is coverage of output types, which reached its widest point with sample-based and compound scoring. The second, and the one occupying every recent release, is making relative skill degrade gracefully: single-model input, comparison groups with one model, and groups missing the requested baseline have each been converted from a cryptic upstream abort into a defined result. That pattern — inherited scoringutils errors being caught and given hub-specific meaning — is the clearest signal of where this package adds value.
◆ Prediction
Expect continued work smoothing scoringutils error surfaces into hub-aware behaviour, and performance attention on relative skill, which was explicitly optimised in the latest release.
mcptools
2.5ANALYTICS
R became a deployable MCP server, not just a local one.
◆ Current state
mcptools lets R act as an MCP server exposing tools to LLM clients, and lets those clients reach into running R sessions. It reached 1.0.0 in mid-2026 after a rename from acquaint and a reversal of its dependency relationship with btw. The trajectory of its releases tracks the MCP specification closely — transport options, protocol version negotiation, and content types have each arrived as the spec settled them.
◆ Where it's heading
The package has moved outward along two axes: transport, from stdio to HTTP to a hosted Posit Connect engine, and content, from text-only tool results to inline images and structured JSON. The most recent release turns to the problems that only appear once something is deployed — choosing the right R session among several, and isolating IPC per user. That shift from capability to multi-user correctness is what a package looks like after it starts being run somewhere other than a developer's laptop.
◆ Prediction
Expect authentication to be addressed on the HTTP transport, which the notes explicitly flag as authless, and continued tracking of MCP protocol revisions as they are published.
Alternatives to hubEvals and mcptools
Other Analytics products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either hubEvals or mcptools.
nabla
nabla dropped its C++ engine to chase exact derivatives at any order.
snowflakeauth
Eight months from first release to keyring caching and workload identity.
kwb.geosalz
A research-project workflow package where the interesting work is in the plumbing.
cyclocomp
A cyclomatic complexity checker that ships once every couple of years, and lands when it does.
revdbayes
Extreme value sampling in pure upkeep mode, mostly answering to Rcpp and CRAN.
See all hubEvals alternatives → · See all mcptools alternatives →
Recent activity from hubEvals and mcptools
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
- 20d agomcptoolsSession selection at tool-call time and per-user IPC
- 24d agohubEvalsScored-forecast counts and faster relative skill
- 1mo agohubEvalsDisaggregated relative skill no longer aborts the whole call
- 1mo agomcptoolsmcptools runs as a Posit Connect R API engine ⚡
- 1mo agohubEvalsSingle-model scoring returns relative skill of 1 instead of erroring
- 5mo agomcptoolsProtocol version negotiation and schema conformance fixes
- 5mo agohubEvalsSample output types and multivariate compound scoring ⚡
- 6mo agohubEvalsScoring on transformed scales via transform arguments
- 9mo agomcptoolsHTTP transport added alongside stdio
- 11mo agohubEvalsFirst release: score_model_out() and the scoringutils bridge
- 11mo agomcptoolsTest fix for an r-devel Fedora clang platform
- 1y agomcptoolsFirst CRAN release, renamed from acquaint with btw dependency reversed
Frequently asked questions
What is the difference between hubEvals and mcptools?
Both compete on the same themes — r-package — within Analytics. hubEvals and mcptools are shipping at a similar cadence (velocity 2.5 vs 2.5, both within Sparkpulse's "active" band). See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Is hubEvals better than mcptools?
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. hubEvals and mcptools are shipping at a similar cadence (velocity 2.5 vs 2.5, both within Sparkpulse's "active" band). For your specific use case, the alternatives sections above list other Analytics products to evaluate alongside.
What are the best alternatives to hubEvals?
Top hubEvals alternatives in Analytics are ranked by recent ship velocity. Browse the "hubEvals alternatives" section above for the current picks, or visit /alternatives/hubevals for the full list with editorial commentary on each.
What are the best alternatives to mcptools?
Top mcptools alternatives in Analytics are ranked by recent ship velocity. Browse the "mcptools alternatives" section above for the current picks, or visit /alternatives/mcptools for the full list with editorial commentary on each.