mlr3proba
mlr3proba is shedding weight as its survival work moves into sibling packages
Comparison · Analytics
probably vs stacks
A side-by-side editorial comparison of probably and stacks — release velocity, themes, recent moves, and the top alternatives to consider.
Shared themes:tidymodels
probably vs stacks: at a glance
What is probably?
The package that made calibration a step instead of an afterthought.
probably started as a small utility for class predictions and equivocal zones, and version 1.0.0 turned it into tidymodels' calibration and uncertainty package: cal_plot_*, cal_estimate_*, cal_validate_* and cal_apply across binary, multiclass and regression problems, plus conformal prediction intervals. Since then the work has been consolidation — a large internal refactor with no API change, split conformal and conformal quantile regression, bound_prediction(), and required_pkgs() and butcher methods so conformal objects can be deployed and stripped.
What is stacks?
Model stacking in tidymodels, quietly migrating off foreach and onto future
stacks builds ensembles from tidymodels tuning results, and its release history is dominated by one long project: replacing foreach-based parallelism with the future framework. That transition completed in 1.1.0, where foreach backends began being ignored with a warning and the minimum R version rose to 4.1. Releases are infrequent and small, with the most recent being a CRAN re-submission rather than a change.
probably vs stacks: editorial side-by-side
probably
0.0ANALYTICS
The package that made calibration a step instead of an afterthought.
◆ Current state
probably started as a small utility for class predictions and equivocal zones, and version 1.0.0 turned it into tidymodels' calibration and uncertainty package: cal_plot_*, cal_estimate_*, cal_validate_* and cal_apply across binary, multiclass and regression problems, plus conformal prediction intervals. Since then the work has been consolidation — a large internal refactor with no API change, split conformal and conformal quantile regression, bound_prediction(), and required_pkgs() and butcher methods so conformal objects can be deployed and stripped.
◆ Where it's heading
The recent releases are about making these objects survive leaving the session. butcher and required_pkgs() methods are what a model needs to be pinned, containerised and served, and their arrival alongside workflows adding a tailor postprocessing stage and vetiver adding probably support points the same way: calibration is being moved out of analysis scripts and into the deployed pipeline. The cal_*_none() reference implementations are the tell that calibration is now something people tune rather than apply once.
◆ Prediction
Expect the calibration functions to be reachable directly from a tuned workflow's postprocessing stage rather than applied to predictions afterwards, following the tailor integration that workflows just shipped.
stacks
0.0ANALYTICS
Model stacking in tidymodels, quietly migrating off foreach and onto future
◆ Current state
stacks builds ensembles from tidymodels tuning results, and its release history is dominated by one long project: replacing foreach-based parallelism with the future framework. That transition completed in 1.1.0, where foreach backends began being ignored with a warning and the minimum R version rose to 4.1. Releases are infrequent and small, with the most recent being a CRAN re-submission rather than a change.
◆ Where it's heading
The package is mature and its remaining work is compatibility rather than capability — tracking the parallelism story across tidymodels, keeping object sizes sane after butchering and reloading, and staying aligned with recipes deprecations. The augment() method added for vetiver compatibility shows the same instinct: fit into the surrounding ecosystem rather than grow independently of it. Nothing in the visible history suggests new ensembling methods are being pursued.
◆ Prediction
With the future migration finished, the next release is most likely maintenance keeping pace with tune and recipes rather than anything users would notice.
Alternatives to probably and stacks
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 probably or stacks.
mlr3viz
mlr3viz keeps the ecosystem's plots working while the plots themselves move out
mlr3tuning
mlr3tuning is rebuilding its async machinery under a stable public surface
timetk
timetk swallowed anomalize whole, then went quiet for two years
modelbased
modelbased is turning marginal effects into a full contrast grammar
parameters
easystats' parameters package absorbs one more model class every few weeks
See all probably alternatives → · See all stacks alternatives →
Recent activity from probably and stacks
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
- 10mo agoprobablyConformal objects gain required_pkgs() and butcher methods
- 1y agoprobablyggplot2 test updates and a clearer validation-set error
- 1y agostacksstacks 1.1.1 re-released to clear a CRAN check note
- 1y agoprobablyCalibration internals refactored; isotonic bootstrap bug fixed
- 1y agostacksstacks 1.1.0 completes the move to future-based parallelism
- 2y agostacksstacks 1.0.5 fixes butchered stack size inflation
- 2y agostacksstacks 1.0.4 introduces future-based parallel processing
- 2y agoprobablyFix grouping sensitivity to variable type
- 2y agostacksstacks 1.0.3 clears recipes deprecations and a type-check bug
- 3y agoprobablySplit conformal and conformal quantile regression added
- 3y agoprobablyCalibration and conformal inference arrive in tidymodels ⚡
- 3y agostacksstacks 1.0.2 adds an augment() method for vetiver compatibility
Frequently asked questions
What is the difference between probably and stacks?
Both compete on the same themes — tidymodels — within Analytics. probably and stacks are shipping at a similar cadence (velocity 0.0 vs 0.0, 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 probably better than stacks?
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. probably and stacks are shipping at a similar cadence (velocity 0.0 vs 0.0, 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 probably?
Top probably alternatives in Analytics are ranked by recent ship velocity. Browse the "probably alternatives" section above for the current picks, or visit /alternatives/probably for the full list with editorial commentary on each.
What are the best alternatives to stacks?
Top stacks alternatives in Analytics are ranked by recent ship velocity. Browse the "stacks alternatives" section above for the current picks, or visit /alternatives/stacks for the full list with editorial commentary on each.