Code Ownership, Measured: The Five-Percent Problem
In short: I asked whether a file with one dominant contributor needs fewer fixes later, once size and activity are already in the model. In all five, it did not help. Telling the model who owns a file never improved its guesses about which files get fixed. The projects did not even agree on a direction. In four of them, files with one heavy contributor were slightly more likely to get fixed later. In the fifth, they were slightly less likely. The five-percent minor-contributor rule could not be tested at all. Too few files had the 21 touches it needs, and those that did mostly shared one outcome. Commit history is a loose stand-in for who understands the code.
Does a file with one clear owner need fewer fixes later than a file many people edit? In five open-source repositories, I could not show that it does.
What ownership is supposed to measure
The idea is about knowledge. A file with one main author probably has someone who understands its design. A file changed by many occasional contributors may have nobody who does. Those contributors can break assumptions only the main author knew about, and bugs follow. Peter Naur called this knowledge the theory of a program. The person who holds it can predict what a change will break, explain why the code looks the way it does, and tell when someone else's change is wrong. That theory lives in people's heads, not in the code.
Nobody can measure that directly, so research measures what Git records: who changed which file. Bird and colleagues did this for compiled Windows binaries at Microsoft. Binaries with many contributors below five percent of the changes had more failures. It is easy to read that as a rule about source files and new contributors. The study measured neither. Later studies moved the threshold, the unit of analysis, and the outcome. So I treated ownership as a hypothesis to test per repository, not as a ready-made risk score.
The stand-in is loose. Two commits in two years make someone the only editor without making them an expert. A reviewer can know a file well and never appear as its author. A formatting commit counts the same as a redesign. Fritz and colleagues tried to close part of this gap. Their degree-of-knowledge model also weighs how recently and how often a developer worked with the code. I used the plain Git version, because that is what the rule everyone repeats is based on.
The link to architecture runs through Conway's law. Ownership shows which people actually work on which parts of the system. A file that many occasional contributors must edit can point to a tangled boundary or a shared hub. It is a people signal that sometimes exposes a structural problem.
How I tested it
Each repository ran through the same four steps.
- Split the history. Ownership comes from the preceding 24-calendar-month training window. The final six calendar months are an untouched holdout.
- Count touches. A touch is one non-merge commit by one contributor to one file. Bots are removed first, and duplicate email addresses are merged.
- Summarise ownership per file. The dominant share is the largest share of touches held by one person. Contributors with at least five percent of a file's touches are major, and the rest are minor. The effective-owner count says how evenly the touches are spread.
- Predict the holdout. For each file, did it appear in a likely-fix commit during those six months? That is the outcome the models try to predict.
Then two models per repository. Each one is a logistic regression: feed it a file's numbers and it returns the probability that the file gets fixed in the holdout. The baseline used only things already known to predict fixes. Those were log LOC, log training touches, prior likely-fix count, lexical complexity, file age, and total contributors. The second model added the dominant share. Both models used files with at least 5 training touches.
If ownership matters, the second model should push the files that later got fixed higher up the list than the baseline does. The gap between them is the result. I report it as a change in PR-AUC, with a bootstrap interval. PR-AUC scores the sort order, not any yes-or-no call. A perfect sort scores one. A random sort scores about the share of files that later needed a fix, which is the prevalence column below. The reported coefficient is standardized: one unit represents one training-set standard deviation in dominant share.
The five-percent rule needs its own model, with the count of minor contributors as the predictor. A contributor with a single touch cannot fall below five percent until the file has at least twenty-one touches, so that model only used files with at least 21 touches. Few files clear that bar. Of the files analysed, webpack has 47 that do, vite 29, prettier 8, eslint 5, and rollup none.
Worked example: how the ranking score is computed
This uses real rows from the published webpack file map. Sort all 251 webpack files in the analysis by their largest one-person share, most spread out first, because that is the order the hypothesis expects to fail more often. Then read down the list. Precision is the share of rows so far that later needed a fix. Recall is the share of all such files found so far. Here are the first eight rows.
| file | largest one-person share | touches | needed a fix | precision so far | recall so far |
|---|---|---|---|---|---|
| lib/json/JsonModulesPlugin.js | 0.286 | 7 | no | 0.000 | 0.000 |
| lib/RuntimeGlobals.js | 0.286 | 7 | yes | 0.500 | 0.007 |
| lib/WebpackOptionsApply.js | 0.289 | 45 | yes | 0.667 | 0.015 |
| lib/DefinePlugin.js | 0.296 | 27 | yes | 0.750 | 0.022 |
| lib/util/internalSerializables.js | 0.300 | 10 | yes | 0.800 | 0.029 |
| lib/ModuleGraph.js | 0.333 | 21 | no | 0.667 | 0.029 |
| lib/EnvironmentPlugin.js | 0.333 | 6 | no | 0.571 | 0.029 |
| lib/ExternalsPlugin.js | 0.333 | 9 | yes | 0.625 | 0.036 |
The score is the area under the precision-recall curve those two columns trace, across all 251 rows. Ranking by spread-out ownership alone gives 0.593. Ranking the other way, most concentrated first, gives 0.492. A random order averages 0.546, the share of these files that later needed a fix. So ownership on its own sorts barely better than chance, and the direction the literature suggests is the better of the two. I computed all three numbers with the same function the analysis uses.
The models in the post are stricter than this. Each repository is scored on several walk-forward folds, and the two models are compared inside each fold. The per-fold gaps are then averaged, weighted by how many files each fold holds. The interval comes from 1000 bootstrap resamples per repository, drawn with a fixed seed of 17 so the numbers reproduce. Each resample redraws the holdout files with replacement, recomputes both areas, and keeps the gap. The interval in the table above is the 2.5th and 97.5th percentile of those 1000 gaps. An interval that spans zero means resampling the same files can flip the sign.
What five repositories showed
| repository | eligible | prevalence | baseline PR-AUC | augmented PR-AUC | delta | interval | dominant share | minor count | ||
|---|---|---|---|---|---|---|---|---|---|---|
| webpack | 251 | 0.545817 | 0.798037 | 0.797751 | -0.000286 | -0.001366 to 0.000775 | 0.036088 | insufficient | ok | 47 |
| eslint | 29 | 0.206897 | 0.359703 | 0.352201 | -0.007503 | -0.117022 to 0.053480 | 0.276285 | insufficient | ok | 5 |
| prettier | 95 | 0.210526 | 0.549515 | 0.556524 | 0.007009 | -0.038076 to 0.055654 | -0.378312 | insufficient | ok | 8 |
| rollup | 47 | 0.085106 | 0.508333 | 0.507258 | -0.001075 | -0.005926 to 0 | 0.043860 | insufficient | ok | 0 |
| vite | 79 | 0.696203 | 0.810068 | 0.794585 | -0.015483 | -0.038445 to 0.006420 | 0.452714 | insufficient | ok | 29 |
Adding ownership did not reliably improve the ranking anywhere. Prettier's ranking improved slightly, and the other four got slightly worse, with Vite dropping the most. Every interval crosses zero, so none of these changes is distinguishable from noise. The dominant-share coefficient also changes sign. It is negative in Prettier and positive in the other four, and strongest in Vite. Concentrated ownership did not point the same way across repositories.
The five-percent model could not be fitted in any repository. Either too few files had twenty-one touches, or the ones that did almost all shared the same outcome. This is the five-percent problem in practice. The rule needs long file histories, and most files in these projects do not have them.
So the hypothesis was not supported here. Five repositories cannot show that ownership is useless, but they do show that its effect did not carry over from one project to the next. One likely reading is that commit history is too loose a stand-in for knowledge. That would be a problem with the measurement, not proof that knowledge is unimportant.
I reran the analysis with the choices changed. Ownership can come from commit touches or from blamed lines, which record who last touched the code that survived. Switching between them changes the coefficient and the ranking delta. Neither one tells you who understands the file. I also tried training windows of 12 and 36 months, a 3-month holdout, touch thresholds of 10 and 40, and minor cutoffs from two to fifty percent. Each change moved the numbers, and some flipped the sign. Many of these variants had too little data to fit a stable model at all. The exact values are in the disclosure below.
All five repositories are JavaScript or TypeScript projects. Commit counts stand in for effort and email addresses for people, and both are rough. Commit subjects stand in for repairs. The audit found 26 of 28 reviewed positive subjects were repairs. That is 92.9% apparent subject-line precision (0.928571), not defect-label precision. The remaining 2 subjects were not repairs. The design measures an association after controls. It cannot show that concentrated contribution causes defects.
What I would do with the signal
I would use it as a review queue, not a score. A concentrated file may be carefully maintained, or it may depend on one person. A diffuse file may be shared well, or it may lack a clear decision maker. Before a risky change, ask the contributors what the history leaves out. Do not rank people, change staffing, or assign blame from these measures.
Every number in this post, and the runs behind it
This table exists for auditing, not for reading. Every value is bound to the checked-in result file, and a build check fails if any of them drifts. Each repository holds its headline run plus the sensitivity runs: training windows of 12, 24 and 36 months, holdouts of 3 and 6 months, touch thresholds of 10, 21 and 40, commit and blame ownership, identity cleaning variants, and minor cutoffs from two to fifty percent. A dash means that run had too little data to fit, not zero.
webpack · 183 values
| repository | variant | claim | value |
|---|---|---|---|
| webpack | training twelve | label | 12-month training |
| webpack | training twelve | group | training window |
| webpack | training twelve | status | ok |
| webpack | training twelve | reason | — |
| webpack | training twelve | coefficient | 0.444691 |
| webpack | training twelve | pr auc delta | 0.001916 |
| webpack | training twelve | baseline pr auc | 0.880047 |
| webpack | training twelve | augmented pr auc | 0.881963 |
| webpack | training twelve | eligible files | 129 |
| webpack | training twelve | minor eligible files | 13 |
| webpack | training twelve | prevalence | 0.697674 |
| webpack | training twelve | minor count available | yes |
| webpack | training twenty-four | label | 24-month training |
| webpack | training twenty-four | group | training window |
| webpack | training twenty-four | status | ok |
| webpack | training twenty-four | reason | — |
| webpack | training twenty-four | coefficient | 0.036088 |
| webpack | training twenty-four | pr auc delta | -0.000286 |
| webpack | training twenty-four | baseline pr auc | 0.798037 |
| webpack | training twenty-four | augmented pr auc | 0.797751 |
| webpack | training twenty-four | eligible files | 251 |
| webpack | training twenty-four | minor eligible files | 47 |
| webpack | training twenty-four | prevalence | 0.545817 |
| webpack | training twenty-four | minor count available | yes |
| webpack | training thirty-six | label | 36-month training |
| webpack | training thirty-six | group | training window |
| webpack | training thirty-six | status | ok |
| webpack | training thirty-six | reason | — |
| webpack | training thirty-six | coefficient | 0.188997 |
| webpack | training thirty-six | pr auc delta | -0.004112 |
| webpack | training thirty-six | baseline pr auc | 0.771134 |
| webpack | training thirty-six | augmented pr auc | 0.767022 |
| webpack | training thirty-six | eligible files | 307 |
| webpack | training thirty-six | minor eligible files | 58 |
| webpack | training thirty-six | prevalence | 0.504886 |
| webpack | training thirty-six | minor count available | yes |
| webpack | touches ten | label | 10 touches |
| webpack | touches ten | group | touch threshold |
| webpack | touches ten | status | ok |
| webpack | touches ten | reason | — |
| webpack | touches ten | coefficient | -0.286604 |
| webpack | touches ten | pr auc delta | 0.002155 |
| webpack | touches ten | baseline pr auc | 0.895003 |
| webpack | touches ten | augmented pr auc | 0.897157 |
| webpack | touches ten | eligible files | 120 |
| webpack | touches ten | minor eligible files | 120 |
| webpack | touches ten | prevalence | 0.716667 |
| webpack | touches ten | minor count available | yes |
| webpack | touches twenty-one | label | 21 touches |
| webpack | touches twenty-one | group | touch threshold |
| webpack | touches twenty-one | status | insufficient |
| webpack | touches twenty-one | reason | training minority class 1 is below 8 fitted parameters |
| webpack | touches twenty-one | coefficient | — |
| webpack | touches twenty-one | pr auc delta | — |
| webpack | touches twenty-one | baseline pr auc | — |
| webpack | touches twenty-one | augmented pr auc | — |
| webpack | touches twenty-one | eligible files | 47 |
| webpack | touches twenty-one | minor eligible files | 47 |
| webpack | touches twenty-one | prevalence | 0.914894 |
| webpack | touches twenty-one | minor count available | yes |
| webpack | touches forty | label | 40 touches |
| webpack | touches forty | group | touch threshold |
| webpack | touches forty | status | insufficient |
| webpack | touches forty | reason | training rows lack both classes |
| webpack | touches forty | coefficient | — |
| webpack | touches forty | pr auc delta | — |
| webpack | touches forty | baseline pr auc | — |
| webpack | touches forty | augmented pr auc | — |
| webpack | touches forty | eligible files | 12 |
| webpack | touches forty | minor eligible files | 12 |
| webpack | touches forty | prevalence | 1 |
| webpack | touches forty | minor count available | yes |
| webpack | holdout six | label | 6-month holdout |
| webpack | holdout six | group | holdout |
| webpack | holdout six | status | ok |
| webpack | holdout six | reason | — |
| webpack | holdout six | coefficient | 0.036088 |
| webpack | holdout six | pr auc delta | -0.000286 |
| webpack | holdout six | baseline pr auc | 0.798037 |
| webpack | holdout six | augmented pr auc | 0.797751 |
| webpack | holdout six | eligible files | 251 |
| webpack | holdout six | minor eligible files | 47 |
| webpack | holdout six | prevalence | 0.545817 |
| webpack | holdout six | minor count available | yes |
| webpack | holdout three | label | 3-month holdout |
| webpack | holdout three | group | holdout |
| webpack | holdout three | status | ok |
| webpack | holdout three | reason | — |
| webpack | holdout three | coefficient | 0.088617 |
| webpack | holdout three | pr auc delta | 0.000673 |
| webpack | holdout three | baseline pr auc | 0.718330 |
| webpack | holdout three | augmented pr auc | 0.719003 |
| webpack | holdout three | eligible files | 255 |
| webpack | holdout three | minor eligible files | 43 |
| webpack | holdout three | prevalence | 0.423529 |
| webpack | holdout three | minor count available | yes |
| webpack | commit | label | commit ownership |
| webpack | commit | group | ownership source |
| webpack | commit | status | ok |
| webpack | commit | reason | — |
| webpack | commit | coefficient | 0.036088 |
| webpack | commit | pr auc delta | -0.000286 |
| webpack | commit | baseline pr auc | 0.798037 |
| webpack | commit | augmented pr auc | 0.797751 |
| webpack | commit | eligible files | 251 |
| webpack | commit | minor eligible files | 47 |
| webpack | commit | prevalence | 0.545817 |
| webpack | commit | minor count available | yes |
| webpack | line | label | line ownership |
| webpack | line | group | ownership source |
| webpack | line | status | ok |
| webpack | line | reason | — |
| webpack | line | coefficient | 0.043366 |
| webpack | line | pr auc delta | -0.000817 |
| webpack | line | baseline pr auc | 0.798037 |
| webpack | line | augmented pr auc | 0.797220 |
| webpack | line | eligible files | 251 |
| webpack | line | minor eligible files | 47 |
| webpack | line | prevalence | 0.545817 |
| webpack | line | minor count available | yes |
| webpack | bots | label | include bots |
| webpack | bots | group | data cleaning |
| webpack | bots | status | ok |
| webpack | bots | reason | — |
| webpack | bots | coefficient | 0.036088 |
| webpack | bots | pr auc delta | -0.000286 |
| webpack | bots | baseline pr auc | 0.798037 |
| webpack | bots | augmented pr auc | 0.797751 |
| webpack | bots | eligible files | 251 |
| webpack | bots | minor eligible files | 47 |
| webpack | bots | prevalence | 0.545817 |
| webpack | bots | minor count available | yes |
| webpack | mailmap | label | without mailmap |
| webpack | mailmap | group | data cleaning |
| webpack | mailmap | status | ok |
| webpack | mailmap | reason | — |
| webpack | mailmap | coefficient | 0.036088 |
| webpack | mailmap | pr auc delta | -0.000286 |
| webpack | mailmap | baseline pr auc | 0.798037 |
| webpack | mailmap | augmented pr auc | 0.797751 |
| webpack | mailmap | eligible files | 251 |
| webpack | mailmap | minor eligible files | 47 |
| webpack | mailmap | prevalence | 0.545817 |
| webpack | mailmap | minor count available | yes |
| webpack | likely fix | label | likely-fix commits |
| webpack | likely fix | group | outcome |
| webpack | likely fix | status | ok |
| webpack | likely fix | reason | — |
| webpack | likely fix | coefficient | 0.036088 |
| webpack | likely fix | pr auc delta | -0.000286 |
| webpack | likely fix | baseline pr auc | 0.798037 |
| webpack | likely fix | augmented pr auc | 0.797751 |
| webpack | likely fix | eligible files | 251 |
| webpack | likely fix | minor eligible files | 47 |
| webpack | likely fix | prevalence | 0.545817 |
| webpack | likely fix | minor count available | yes |
| webpack | all change | label | all changes |
| webpack | all change | group | outcome |
| webpack | all change | status | ok |
| webpack | all change | reason | — |
| webpack | all change | coefficient | -0.075075 |
| webpack | all change | pr auc delta | 0.000282 |
| webpack | all change | baseline pr auc | 0.957857 |
| webpack | all change | augmented pr auc | 0.958139 |
| webpack | all change | eligible files | 251 |
| webpack | all change | minor eligible files | 47 |
| webpack | all change | prevalence | 0.888446 |
| webpack | all change | minor count available | yes |
| webpack | minor cutoff two percent | available | yes |
| webpack | minor cutoff two percent | status | insufficient |
| webpack | minor cutoff two percent | reason | training minority class 1 is below 8 fitted parameters |
| webpack | minor cutoff five percent | available | yes |
| webpack | minor cutoff five percent | status | insufficient |
| webpack | minor cutoff five percent | reason | training minority class 1 is below 8 fitted parameters |
| webpack | minor cutoff ten percent | available | yes |
| webpack | minor cutoff ten percent | status | insufficient |
| webpack | minor cutoff ten percent | reason | training minority class 1 is below 8 fitted parameters |
| webpack | minor cutoff twenty percent | available | yes |
| webpack | minor cutoff twenty percent | status | insufficient |
| webpack | minor cutoff twenty percent | reason | training minority class 1 is below 8 fitted parameters |
| webpack | minor cutoff fifty percent | available | yes |
| webpack | minor cutoff fifty percent | status | insufficient |
| webpack | minor cutoff fifty percent | reason | training minority class 1 is below 8 fitted parameters |
eslint · 183 values
| repository | variant | claim | value |
|---|---|---|---|
| eslint | training twelve | label | 12-month training |
| eslint | training twelve | group | training window |
| eslint | training twelve | status | insufficient |
| eslint | training twelve | reason | training rows 32 are below 40 required for 8 fitted parameters |
| eslint | training twelve | coefficient | — |
| eslint | training twelve | pr auc delta | — |
| eslint | training twelve | baseline pr auc | — |
| eslint | training twelve | augmented pr auc | — |
| eslint | training twelve | eligible files | 17 |
| eslint | training twelve | minor eligible files | 1 |
| eslint | training twelve | prevalence | 0.294118 |
| eslint | training twelve | minor count available | yes |
| eslint | training twenty-four | label | 24-month training |
| eslint | training twenty-four | group | training window |
| eslint | training twenty-four | status | ok |
| eslint | training twenty-four | reason | — |
| eslint | training twenty-four | coefficient | 0.276285 |
| eslint | training twenty-four | pr auc delta | -0.007503 |
| eslint | training twenty-four | baseline pr auc | 0.359703 |
| eslint | training twenty-four | augmented pr auc | 0.352201 |
| eslint | training twenty-four | eligible files | 29 |
| eslint | training twenty-four | minor eligible files | 5 |
| eslint | training twenty-four | prevalence | 0.206897 |
| eslint | training twenty-four | minor count available | yes |
| eslint | training thirty-six | label | 36-month training |
| eslint | training thirty-six | group | training window |
| eslint | training thirty-six | status | ok |
| eslint | training thirty-six | reason | — |
| eslint | training thirty-six | coefficient | 0.328153 |
| eslint | training thirty-six | pr auc delta | -0.002282 |
| eslint | training thirty-six | baseline pr auc | 0.283631 |
| eslint | training thirty-six | augmented pr auc | 0.281349 |
| eslint | training thirty-six | eligible files | 112 |
| eslint | training thirty-six | minor eligible files | 7 |
| eslint | training thirty-six | prevalence | 0.133929 |
| eslint | training thirty-six | minor count available | yes |
| eslint | touches ten | label | 10 touches |
| eslint | touches ten | group | touch threshold |
| eslint | touches ten | status | insufficient |
| eslint | touches ten | reason | training rows 26 are below 40 required for 8 fitted parameters |
| eslint | touches ten | coefficient | — |
| eslint | touches ten | pr auc delta | — |
| eslint | touches ten | baseline pr auc | — |
| eslint | touches ten | augmented pr auc | — |
| eslint | touches ten | eligible files | 12 |
| eslint | touches ten | minor eligible files | 12 |
| eslint | touches ten | prevalence | 0.333333 |
| eslint | touches ten | minor count available | yes |
| eslint | touches twenty-one | label | 21 touches |
| eslint | touches twenty-one | group | touch threshold |
| eslint | touches twenty-one | status | insufficient |
| eslint | touches twenty-one | reason | training rows 11 are below 40 required for 8 fitted parameters |
| eslint | touches twenty-one | coefficient | — |
| eslint | touches twenty-one | pr auc delta | — |
| eslint | touches twenty-one | baseline pr auc | — |
| eslint | touches twenty-one | augmented pr auc | — |
| eslint | touches twenty-one | eligible files | 5 |
| eslint | touches twenty-one | minor eligible files | 5 |
| eslint | touches twenty-one | prevalence | 0.400000 |
| eslint | touches twenty-one | minor count available | yes |
| eslint | touches forty | label | 40 touches |
| eslint | touches forty | group | touch threshold |
| eslint | touches forty | status | insufficient |
| eslint | touches forty | reason | training rows lack both classes |
| eslint | touches forty | coefficient | — |
| eslint | touches forty | pr auc delta | — |
| eslint | touches forty | baseline pr auc | — |
| eslint | touches forty | augmented pr auc | — |
| eslint | touches forty | eligible files | 2 |
| eslint | touches forty | minor eligible files | 2 |
| eslint | touches forty | prevalence | 0.500000 |
| eslint | touches forty | minor count available | yes |
| eslint | holdout six | label | 6-month holdout |
| eslint | holdout six | group | holdout |
| eslint | holdout six | status | ok |
| eslint | holdout six | reason | — |
| eslint | holdout six | coefficient | 0.276285 |
| eslint | holdout six | pr auc delta | -0.007503 |
| eslint | holdout six | baseline pr auc | 0.359703 |
| eslint | holdout six | augmented pr auc | 0.352201 |
| eslint | holdout six | eligible files | 29 |
| eslint | holdout six | minor eligible files | 5 |
| eslint | holdout six | prevalence | 0.206897 |
| eslint | holdout six | minor count available | yes |
| eslint | holdout three | label | 3-month holdout |
| eslint | holdout three | group | holdout |
| eslint | holdout three | status | ok |
| eslint | holdout three | reason | — |
| eslint | holdout three | coefficient | 0.681578 |
| eslint | holdout three | pr auc delta | 0.133548 |
| eslint | holdout three | baseline pr auc | 0.304721 |
| eslint | holdout three | augmented pr auc | 0.438268 |
| eslint | holdout three | eligible files | 30 |
| eslint | holdout three | minor eligible files | 5 |
| eslint | holdout three | prevalence | 0.166667 |
| eslint | holdout three | minor count available | yes |
| eslint | commit | label | commit ownership |
| eslint | commit | group | ownership source |
| eslint | commit | status | ok |
| eslint | commit | reason | — |
| eslint | commit | coefficient | 0.276285 |
| eslint | commit | pr auc delta | -0.007503 |
| eslint | commit | baseline pr auc | 0.359703 |
| eslint | commit | augmented pr auc | 0.352201 |
| eslint | commit | eligible files | 29 |
| eslint | commit | minor eligible files | 5 |
| eslint | commit | prevalence | 0.206897 |
| eslint | commit | minor count available | yes |
| eslint | line | label | line ownership |
| eslint | line | group | ownership source |
| eslint | line | status | ok |
| eslint | line | reason | — |
| eslint | line | coefficient | 0.067121 |
| eslint | line | pr auc delta | 0.010504 |
| eslint | line | baseline pr auc | 0.359703 |
| eslint | line | augmented pr auc | 0.370208 |
| eslint | line | eligible files | 29 |
| eslint | line | minor eligible files | 5 |
| eslint | line | prevalence | 0.206897 |
| eslint | line | minor count available | yes |
| eslint | bots | label | include bots |
| eslint | bots | group | data cleaning |
| eslint | bots | status | ok |
| eslint | bots | reason | — |
| eslint | bots | coefficient | 0.276285 |
| eslint | bots | pr auc delta | -0.007503 |
| eslint | bots | baseline pr auc | 0.359703 |
| eslint | bots | augmented pr auc | 0.352201 |
| eslint | bots | eligible files | 29 |
| eslint | bots | minor eligible files | 5 |
| eslint | bots | prevalence | 0.206897 |
| eslint | bots | minor count available | yes |
| eslint | mailmap | label | without mailmap |
| eslint | mailmap | group | data cleaning |
| eslint | mailmap | status | ok |
| eslint | mailmap | reason | — |
| eslint | mailmap | coefficient | 0.276285 |
| eslint | mailmap | pr auc delta | -0.007503 |
| eslint | mailmap | baseline pr auc | 0.359703 |
| eslint | mailmap | augmented pr auc | 0.352201 |
| eslint | mailmap | eligible files | 29 |
| eslint | mailmap | minor eligible files | 5 |
| eslint | mailmap | prevalence | 0.206897 |
| eslint | mailmap | minor count available | yes |
| eslint | likely fix | label | likely-fix commits |
| eslint | likely fix | group | outcome |
| eslint | likely fix | status | ok |
| eslint | likely fix | reason | — |
| eslint | likely fix | coefficient | 0.276285 |
| eslint | likely fix | pr auc delta | -0.007503 |
| eslint | likely fix | baseline pr auc | 0.359703 |
| eslint | likely fix | augmented pr auc | 0.352201 |
| eslint | likely fix | eligible files | 29 |
| eslint | likely fix | minor eligible files | 5 |
| eslint | likely fix | prevalence | 0.206897 |
| eslint | likely fix | minor count available | yes |
| eslint | all change | label | all changes |
| eslint | all change | group | outcome |
| eslint | all change | status | ok |
| eslint | all change | reason | — |
| eslint | all change | coefficient | 0.159696 |
| eslint | all change | pr auc delta | -0.001277 |
| eslint | all change | baseline pr auc | 0.724452 |
| eslint | all change | augmented pr auc | 0.723175 |
| eslint | all change | eligible files | 29 |
| eslint | all change | minor eligible files | 5 |
| eslint | all change | prevalence | 0.517241 |
| eslint | all change | minor count available | yes |
| eslint | minor cutoff two percent | available | yes |
| eslint | minor cutoff two percent | status | insufficient |
| eslint | minor cutoff two percent | reason | training rows 11 are below 40 required for 8 fitted parameters |
| eslint | minor cutoff five percent | available | yes |
| eslint | minor cutoff five percent | status | insufficient |
| eslint | minor cutoff five percent | reason | training rows 11 are below 40 required for 8 fitted parameters |
| eslint | minor cutoff ten percent | available | yes |
| eslint | minor cutoff ten percent | status | insufficient |
| eslint | minor cutoff ten percent | reason | training rows 11 are below 40 required for 8 fitted parameters |
| eslint | minor cutoff twenty percent | available | yes |
| eslint | minor cutoff twenty percent | status | insufficient |
| eslint | minor cutoff twenty percent | reason | training rows 11 are below 40 required for 8 fitted parameters |
| eslint | minor cutoff fifty percent | available | yes |
| eslint | minor cutoff fifty percent | status | insufficient |
| eslint | minor cutoff fifty percent | reason | training rows 11 are below 40 required for 8 fitted parameters |
prettier · 183 values
| repository | variant | claim | value |
|---|---|---|---|
| prettier | training twelve | label | 12-month training |
| prettier | training twelve | group | training window |
| prettier | training twelve | status | ok |
| prettier | training twelve | reason | — |
| prettier | training twelve | coefficient | -0.056489 |
| prettier | training twelve | pr auc delta | 0.003159 |
| prettier | training twelve | baseline pr auc | 0.557423 |
| prettier | training twelve | augmented pr auc | 0.560582 |
| prettier | training twelve | eligible files | 74 |
| prettier | training twelve | minor eligible files | 2 |
| prettier | training twelve | prevalence | 0.229730 |
| prettier | training twelve | minor count available | yes |
| prettier | training twenty-four | label | 24-month training |
| prettier | training twenty-four | group | training window |
| prettier | training twenty-four | status | ok |
| prettier | training twenty-four | reason | — |
| prettier | training twenty-four | coefficient | -0.378312 |
| prettier | training twenty-four | pr auc delta | 0.007009 |
| prettier | training twenty-four | baseline pr auc | 0.549515 |
| prettier | training twenty-four | augmented pr auc | 0.556524 |
| prettier | training twenty-four | eligible files | 95 |
| prettier | training twenty-four | minor eligible files | 8 |
| prettier | training twenty-four | prevalence | 0.210526 |
| prettier | training twenty-four | minor count available | yes |
| prettier | training thirty-six | label | 36-month training |
| prettier | training thirty-six | group | training window |
| prettier | training thirty-six | status | ok |
| prettier | training thirty-six | reason | — |
| prettier | training thirty-six | coefficient | 0.184680 |
| prettier | training thirty-six | pr auc delta | -0.046072 |
| prettier | training thirty-six | baseline pr auc | 0.527554 |
| prettier | training thirty-six | augmented pr auc | 0.481482 |
| prettier | training thirty-six | eligible files | 147 |
| prettier | training thirty-six | minor eligible files | 26 |
| prettier | training thirty-six | prevalence | 0.163265 |
| prettier | training thirty-six | minor count available | yes |
| prettier | touches ten | label | 10 touches |
| prettier | touches ten | group | touch threshold |
| prettier | touches ten | status | ok |
| prettier | touches ten | reason | — |
| prettier | touches ten | coefficient | -0.392087 |
| prettier | touches ten | pr auc delta | 0.023285 |
| prettier | touches ten | baseline pr auc | 0.562882 |
| prettier | touches ten | augmented pr auc | 0.586167 |
| prettier | touches ten | eligible files | 47 |
| prettier | touches ten | minor eligible files | 47 |
| prettier | touches ten | prevalence | 0.255319 |
| prettier | touches ten | minor count available | yes |
| prettier | touches twenty-one | label | 21 touches |
| prettier | touches twenty-one | group | touch threshold |
| prettier | touches twenty-one | status | insufficient |
| prettier | touches twenty-one | reason | training rows 12 are below 40 required for 8 fitted parameters |
| prettier | touches twenty-one | coefficient | — |
| prettier | touches twenty-one | pr auc delta | — |
| prettier | touches twenty-one | baseline pr auc | — |
| prettier | touches twenty-one | augmented pr auc | — |
| prettier | touches twenty-one | eligible files | 8 |
| prettier | touches twenty-one | minor eligible files | 8 |
| prettier | touches twenty-one | prevalence | 0.250000 |
| prettier | touches twenty-one | minor count available | yes |
| prettier | touches forty | label | 40 touches |
| prettier | touches forty | group | touch threshold |
| prettier | touches forty | status | insufficient |
| prettier | touches forty | reason | training rows lack both classes |
| prettier | touches forty | coefficient | — |
| prettier | touches forty | pr auc delta | — |
| prettier | touches forty | baseline pr auc | — |
| prettier | touches forty | augmented pr auc | — |
| prettier | touches forty | eligible files | 1 |
| prettier | touches forty | minor eligible files | 1 |
| prettier | touches forty | prevalence | 1 |
| prettier | touches forty | minor count available | yes |
| prettier | holdout six | label | 6-month holdout |
| prettier | holdout six | group | holdout |
| prettier | holdout six | status | ok |
| prettier | holdout six | reason | — |
| prettier | holdout six | coefficient | -0.378312 |
| prettier | holdout six | pr auc delta | 0.007009 |
| prettier | holdout six | baseline pr auc | 0.549515 |
| prettier | holdout six | augmented pr auc | 0.556524 |
| prettier | holdout six | eligible files | 95 |
| prettier | holdout six | minor eligible files | 8 |
| prettier | holdout six | prevalence | 0.210526 |
| prettier | holdout six | minor count available | yes |
| prettier | holdout three | label | 3-month holdout |
| prettier | holdout three | group | holdout |
| prettier | holdout three | status | ok |
| prettier | holdout three | reason | — |
| prettier | holdout three | coefficient | 0.230570 |
| prettier | holdout three | pr auc delta | 0.000223 |
| prettier | holdout three | baseline pr auc | 0.540332 |
| prettier | holdout three | augmented pr auc | 0.540555 |
| prettier | holdout three | eligible files | 122 |
| prettier | holdout three | minor eligible files | 13 |
| prettier | holdout three | prevalence | 0.163934 |
| prettier | holdout three | minor count available | yes |
| prettier | commit | label | commit ownership |
| prettier | commit | group | ownership source |
| prettier | commit | status | ok |
| prettier | commit | reason | — |
| prettier | commit | coefficient | -0.378312 |
| prettier | commit | pr auc delta | 0.007009 |
| prettier | commit | baseline pr auc | 0.549515 |
| prettier | commit | augmented pr auc | 0.556524 |
| prettier | commit | eligible files | 95 |
| prettier | commit | minor eligible files | 8 |
| prettier | commit | prevalence | 0.210526 |
| prettier | commit | minor count available | yes |
| prettier | line | label | line ownership |
| prettier | line | group | ownership source |
| prettier | line | status | ok |
| prettier | line | reason | — |
| prettier | line | coefficient | -0.381293 |
| prettier | line | pr auc delta | 0.076276 |
| prettier | line | baseline pr auc | 0.549515 |
| prettier | line | augmented pr auc | 0.625790 |
| prettier | line | eligible files | 95 |
| prettier | line | minor eligible files | 8 |
| prettier | line | prevalence | 0.210526 |
| prettier | line | minor count available | yes |
| prettier | bots | label | include bots |
| prettier | bots | group | data cleaning |
| prettier | bots | status | ok |
| prettier | bots | reason | — |
| prettier | bots | coefficient | -0.311133 |
| prettier | bots | pr auc delta | 0.024569 |
| prettier | bots | baseline pr auc | 0.525417 |
| prettier | bots | augmented pr auc | 0.549986 |
| prettier | bots | eligible files | 103 |
| prettier | bots | minor eligible files | 8 |
| prettier | bots | prevalence | 0.213592 |
| prettier | bots | minor count available | yes |
| prettier | mailmap | label | without mailmap |
| prettier | mailmap | group | data cleaning |
| prettier | mailmap | status | ok |
| prettier | mailmap | reason | — |
| prettier | mailmap | coefficient | -0.378312 |
| prettier | mailmap | pr auc delta | 0.007009 |
| prettier | mailmap | baseline pr auc | 0.549515 |
| prettier | mailmap | augmented pr auc | 0.556524 |
| prettier | mailmap | eligible files | 95 |
| prettier | mailmap | minor eligible files | 8 |
| prettier | mailmap | prevalence | 0.210526 |
| prettier | mailmap | minor count available | yes |
| prettier | likely fix | label | likely-fix commits |
| prettier | likely fix | group | outcome |
| prettier | likely fix | status | ok |
| prettier | likely fix | reason | — |
| prettier | likely fix | coefficient | -0.378312 |
| prettier | likely fix | pr auc delta | 0.007009 |
| prettier | likely fix | baseline pr auc | 0.549515 |
| prettier | likely fix | augmented pr auc | 0.556524 |
| prettier | likely fix | eligible files | 95 |
| prettier | likely fix | minor eligible files | 8 |
| prettier | likely fix | prevalence | 0.210526 |
| prettier | likely fix | minor count available | yes |
| prettier | all change | label | all changes |
| prettier | all change | group | outcome |
| prettier | all change | status | ok |
| prettier | all change | reason | — |
| prettier | all change | coefficient | 0.180660 |
| prettier | all change | pr auc delta | -0.000948 |
| prettier | all change | baseline pr auc | 0.944800 |
| prettier | all change | augmented pr auc | 0.943853 |
| prettier | all change | eligible files | 95 |
| prettier | all change | minor eligible files | 8 |
| prettier | all change | prevalence | 0.789474 |
| prettier | all change | minor count available | yes |
| prettier | minor cutoff two percent | available | no |
| prettier | minor cutoff two percent | status | unavailable |
| prettier | minor cutoff two percent | reason | no observed below-cutoff owner in training or test rows |
| prettier | minor cutoff five percent | available | yes |
| prettier | minor cutoff five percent | status | insufficient |
| prettier | minor cutoff five percent | reason | training rows 12 are below 40 required for 8 fitted parameters |
| prettier | minor cutoff ten percent | available | yes |
| prettier | minor cutoff ten percent | status | insufficient |
| prettier | minor cutoff ten percent | reason | training rows 12 are below 40 required for 8 fitted parameters |
| prettier | minor cutoff twenty percent | available | yes |
| prettier | minor cutoff twenty percent | status | insufficient |
| prettier | minor cutoff twenty percent | reason | training rows 12 are below 40 required for 8 fitted parameters |
| prettier | minor cutoff fifty percent | available | yes |
| prettier | minor cutoff fifty percent | status | insufficient |
| prettier | minor cutoff fifty percent | reason | training rows 12 are below 40 required for 8 fitted parameters |
rollup · 183 values
| repository | variant | claim | value |
|---|---|---|---|
| rollup | training twelve | label | 12-month training |
| rollup | training twelve | group | training window |
| rollup | training twelve | status | ok |
| rollup | training twelve | reason | — |
| rollup | training twelve | coefficient | -1.054052 |
| rollup | training twelve | pr auc delta | 0.040079 |
| rollup | training twelve | baseline pr auc | 0.319444 |
| rollup | training twelve | augmented pr auc | 0.359524 |
| rollup | training twelve | eligible files | 9 |
| rollup | training twelve | minor eligible files | 0 |
| rollup | training twelve | prevalence | 0.333333 |
| rollup | training twelve | minor count available | no |
| rollup | training twenty-four | label | 24-month training |
| rollup | training twenty-four | group | training window |
| rollup | training twenty-four | status | ok |
| rollup | training twenty-four | reason | — |
| rollup | training twenty-four | coefficient | 0.043860 |
| rollup | training twenty-four | pr auc delta | -0.001075 |
| rollup | training twenty-four | baseline pr auc | 0.508333 |
| rollup | training twenty-four | augmented pr auc | 0.507258 |
| rollup | training twenty-four | eligible files | 47 |
| rollup | training twenty-four | minor eligible files | 0 |
| rollup | training twenty-four | prevalence | 0.085106 |
| rollup | training twenty-four | minor count available | yes |
| rollup | training thirty-six | label | 36-month training |
| rollup | training thirty-six | group | training window |
| rollup | training thirty-six | status | ok |
| rollup | training thirty-six | reason | — |
| rollup | training thirty-six | coefficient | 0.242093 |
| rollup | training thirty-six | pr auc delta | -0.036574 |
| rollup | training thirty-six | baseline pr auc | 0.445185 |
| rollup | training thirty-six | augmented pr auc | 0.408611 |
| rollup | training thirty-six | eligible files | 95 |
| rollup | training thirty-six | minor eligible files | 4 |
| rollup | training thirty-six | prevalence | 0.073684 |
| rollup | training thirty-six | minor count available | yes |
| rollup | touches ten | label | 10 touches |
| rollup | touches ten | group | touch threshold |
| rollup | touches ten | status | ok |
| rollup | touches ten | reason | — |
| rollup | touches ten | coefficient | 0.466261 |
| rollup | touches ten | pr auc delta | -0.023810 |
| rollup | touches ten | baseline pr auc | 0.333333 |
| rollup | touches ten | augmented pr auc | 0.309524 |
| rollup | touches ten | eligible files | 11 |
| rollup | touches ten | minor eligible files | 11 |
| rollup | touches ten | prevalence | 0.181818 |
| rollup | touches ten | minor count available | yes |
| rollup | touches twenty-one | label | 21 touches |
| rollup | touches twenty-one | group | touch threshold |
| rollup | touches twenty-one | status | insufficient |
| rollup | touches twenty-one | reason | test rows lack both classes |
| rollup | touches twenty-one | coefficient | — |
| rollup | touches twenty-one | pr auc delta | — |
| rollup | touches twenty-one | baseline pr auc | — |
| rollup | touches twenty-one | augmented pr auc | — |
| rollup | touches twenty-one | eligible files | 0 |
| rollup | touches twenty-one | minor eligible files | 0 |
| rollup | touches twenty-one | prevalence | — |
| rollup | touches twenty-one | minor count available | yes |
| rollup | touches forty | label | 40 touches |
| rollup | touches forty | group | touch threshold |
| rollup | touches forty | status | insufficient |
| rollup | touches forty | reason | training rows lack both classes |
| rollup | touches forty | coefficient | — |
| rollup | touches forty | pr auc delta | — |
| rollup | touches forty | baseline pr auc | — |
| rollup | touches forty | augmented pr auc | — |
| rollup | touches forty | eligible files | 0 |
| rollup | touches forty | minor eligible files | 0 |
| rollup | touches forty | prevalence | — |
| rollup | touches forty | minor count available | no |
| rollup | holdout six | label | 6-month holdout |
| rollup | holdout six | group | holdout |
| rollup | holdout six | status | ok |
| rollup | holdout six | reason | — |
| rollup | holdout six | coefficient | 0.043860 |
| rollup | holdout six | pr auc delta | -0.001075 |
| rollup | holdout six | baseline pr auc | 0.508333 |
| rollup | holdout six | augmented pr auc | 0.507258 |
| rollup | holdout six | eligible files | 47 |
| rollup | holdout six | minor eligible files | 0 |
| rollup | holdout six | prevalence | 0.085106 |
| rollup | holdout six | minor count available | yes |
| rollup | holdout three | label | 3-month holdout |
| rollup | holdout three | group | holdout |
| rollup | holdout three | status | insufficient |
| rollup | holdout three | reason | test rows lack both classes |
| rollup | holdout three | coefficient | — |
| rollup | holdout three | pr auc delta | — |
| rollup | holdout three | baseline pr auc | — |
| rollup | holdout three | augmented pr auc | — |
| rollup | holdout three | eligible files | 42 |
| rollup | holdout three | minor eligible files | 0 |
| rollup | holdout three | prevalence | 0 |
| rollup | holdout three | minor count available | no |
| rollup | commit | label | commit ownership |
| rollup | commit | group | ownership source |
| rollup | commit | status | ok |
| rollup | commit | reason | — |
| rollup | commit | coefficient | 0.043860 |
| rollup | commit | pr auc delta | -0.001075 |
| rollup | commit | baseline pr auc | 0.508333 |
| rollup | commit | augmented pr auc | 0.507258 |
| rollup | commit | eligible files | 47 |
| rollup | commit | minor eligible files | 0 |
| rollup | commit | prevalence | 0.085106 |
| rollup | commit | minor count available | yes |
| rollup | line | label | line ownership |
| rollup | line | group | ownership source |
| rollup | line | status | ok |
| rollup | line | reason | — |
| rollup | line | coefficient | -0.283960 |
| rollup | line | pr auc delta | 0 |
| rollup | line | baseline pr auc | 0.508333 |
| rollup | line | augmented pr auc | 0.508333 |
| rollup | line | eligible files | 47 |
| rollup | line | minor eligible files | 0 |
| rollup | line | prevalence | 0.085106 |
| rollup | line | minor count available | yes |
| rollup | bots | label | include bots |
| rollup | bots | group | data cleaning |
| rollup | bots | status | ok |
| rollup | bots | reason | — |
| rollup | bots | coefficient | -0.017445 |
| rollup | bots | pr auc delta | -0.002179 |
| rollup | bots | baseline pr auc | 0.402865 |
| rollup | bots | augmented pr auc | 0.400686 |
| rollup | bots | eligible files | 55 |
| rollup | bots | minor eligible files | 0 |
| rollup | bots | prevalence | 0.109091 |
| rollup | bots | minor count available | yes |
| rollup | mailmap | label | without mailmap |
| rollup | mailmap | group | data cleaning |
| rollup | mailmap | status | ok |
| rollup | mailmap | reason | — |
| rollup | mailmap | coefficient | 0.043860 |
| rollup | mailmap | pr auc delta | -0.001075 |
| rollup | mailmap | baseline pr auc | 0.508333 |
| rollup | mailmap | augmented pr auc | 0.507258 |
| rollup | mailmap | eligible files | 47 |
| rollup | mailmap | minor eligible files | 0 |
| rollup | mailmap | prevalence | 0.085106 |
| rollup | mailmap | minor count available | yes |
| rollup | likely fix | label | likely-fix commits |
| rollup | likely fix | group | outcome |
| rollup | likely fix | status | ok |
| rollup | likely fix | reason | — |
| rollup | likely fix | coefficient | 0.043860 |
| rollup | likely fix | pr auc delta | -0.001075 |
| rollup | likely fix | baseline pr auc | 0.508333 |
| rollup | likely fix | augmented pr auc | 0.507258 |
| rollup | likely fix | eligible files | 47 |
| rollup | likely fix | minor eligible files | 0 |
| rollup | likely fix | prevalence | 0.085106 |
| rollup | likely fix | minor count available | yes |
| rollup | all change | label | all changes |
| rollup | all change | group | outcome |
| rollup | all change | status | ok |
| rollup | all change | reason | — |
| rollup | all change | coefficient | -0.030657 |
| rollup | all change | pr auc delta | 0.001164 |
| rollup | all change | baseline pr auc | 0.691039 |
| rollup | all change | augmented pr auc | 0.692203 |
| rollup | all change | eligible files | 47 |
| rollup | all change | minor eligible files | 0 |
| rollup | all change | prevalence | 0.382979 |
| rollup | all change | minor count available | yes |
| rollup | minor cutoff two percent | available | no |
| rollup | minor cutoff two percent | status | unavailable |
| rollup | minor cutoff two percent | reason | no observed below-cutoff owner in training or test rows |
| rollup | minor cutoff five percent | available | yes |
| rollup | minor cutoff five percent | status | insufficient |
| rollup | minor cutoff five percent | reason | test rows lack both classes |
| rollup | minor cutoff ten percent | available | yes |
| rollup | minor cutoff ten percent | status | insufficient |
| rollup | minor cutoff ten percent | reason | test rows lack both classes |
| rollup | minor cutoff twenty percent | available | yes |
| rollup | minor cutoff twenty percent | status | insufficient |
| rollup | minor cutoff twenty percent | reason | test rows lack both classes |
| rollup | minor cutoff fifty percent | available | yes |
| rollup | minor cutoff fifty percent | status | insufficient |
| rollup | minor cutoff fifty percent | reason | test rows lack both classes |
vite · 183 values
| repository | variant | claim | value |
|---|---|---|---|
| vite | training twelve | label | 12-month training |
| vite | training twelve | group | training window |
| vite | training twelve | status | ok |
| vite | training twelve | reason | — |
| vite | training twelve | coefficient | 0.532056 |
| vite | training twelve | pr auc delta | -0.003540 |
| vite | training twelve | baseline pr auc | 0.853439 |
| vite | training twelve | augmented pr auc | 0.849899 |
| vite | training twelve | eligible files | 50 |
| vite | training twelve | minor eligible files | 15 |
| vite | training twelve | prevalence | 0.780000 |
| vite | training twelve | minor count available | yes |
| vite | training twenty-four | label | 24-month training |
| vite | training twenty-four | group | training window |
| vite | training twenty-four | status | ok |
| vite | training twenty-four | reason | — |
| vite | training twenty-four | coefficient | 0.452714 |
| vite | training twenty-four | pr auc delta | -0.015483 |
| vite | training twenty-four | baseline pr auc | 0.810068 |
| vite | training twenty-four | augmented pr auc | 0.794585 |
| vite | training twenty-four | eligible files | 79 |
| vite | training twenty-four | minor eligible files | 29 |
| vite | training twenty-four | prevalence | 0.696203 |
| vite | training twenty-four | minor count available | yes |
| vite | training thirty-six | label | 36-month training |
| vite | training thirty-six | group | training window |
| vite | training thirty-six | status | ok |
| vite | training thirty-six | reason | — |
| vite | training thirty-six | coefficient | 0.191272 |
| vite | training thirty-six | pr auc delta | -0.005634 |
| vite | training thirty-six | baseline pr auc | 0.830090 |
| vite | training thirty-six | augmented pr auc | 0.824457 |
| vite | training thirty-six | eligible files | 89 |
| vite | training thirty-six | minor eligible files | 43 |
| vite | training thirty-six | prevalence | 0.651685 |
| vite | training thirty-six | minor count available | yes |
| vite | touches ten | label | 10 touches |
| vite | touches ten | group | touch threshold |
| vite | touches ten | status | ok |
| vite | touches ten | reason | — |
| vite | touches ten | coefficient | 1.121782 |
| vite | touches ten | pr auc delta | -0.039060 |
| vite | touches ten | baseline pr auc | 0.840150 |
| vite | touches ten | augmented pr auc | 0.801090 |
| vite | touches ten | eligible files | 54 |
| vite | touches ten | minor eligible files | 54 |
| vite | touches ten | prevalence | 0.777778 |
| vite | touches ten | minor count available | yes |
| vite | touches twenty-one | label | 21 touches |
| vite | touches twenty-one | group | touch threshold |
| vite | touches twenty-one | status | insufficient |
| vite | touches twenty-one | reason | training minority class 6 is below 8 fitted parameters |
| vite | touches twenty-one | coefficient | — |
| vite | touches twenty-one | pr auc delta | — |
| vite | touches twenty-one | baseline pr auc | — |
| vite | touches twenty-one | augmented pr auc | — |
| vite | touches twenty-one | eligible files | 29 |
| vite | touches twenty-one | minor eligible files | 29 |
| vite | touches twenty-one | prevalence | 0.862069 |
| vite | touches twenty-one | minor count available | yes |
| vite | touches forty | label | 40 touches |
| vite | touches forty | group | touch threshold |
| vite | touches forty | status | insufficient |
| vite | touches forty | reason | training rows lack both classes |
| vite | touches forty | coefficient | — |
| vite | touches forty | pr auc delta | — |
| vite | touches forty | baseline pr auc | — |
| vite | touches forty | augmented pr auc | — |
| vite | touches forty | eligible files | 12 |
| vite | touches forty | minor eligible files | 12 |
| vite | touches forty | prevalence | 1 |
| vite | touches forty | minor count available | yes |
| vite | holdout six | label | 6-month holdout |
| vite | holdout six | group | holdout |
| vite | holdout six | status | ok |
| vite | holdout six | reason | — |
| vite | holdout six | coefficient | 0.452714 |
| vite | holdout six | pr auc delta | -0.015483 |
| vite | holdout six | baseline pr auc | 0.810068 |
| vite | holdout six | augmented pr auc | 0.794585 |
| vite | holdout six | eligible files | 79 |
| vite | holdout six | minor eligible files | 29 |
| vite | holdout six | prevalence | 0.696203 |
| vite | holdout six | minor count available | yes |
| vite | holdout three | label | 3-month holdout |
| vite | holdout three | group | holdout |
| vite | holdout three | status | ok |
| vite | holdout three | reason | — |
| vite | holdout three | coefficient | 0.025438 |
| vite | holdout three | pr auc delta | -0.001333 |
| vite | holdout three | baseline pr auc | 0.756117 |
| vite | holdout three | augmented pr auc | 0.754784 |
| vite | holdout three | eligible files | 83 |
| vite | holdout three | minor eligible files | 30 |
| vite | holdout three | prevalence | 0.506024 |
| vite | holdout three | minor count available | yes |
| vite | commit | label | commit ownership |
| vite | commit | group | ownership source |
| vite | commit | status | ok |
| vite | commit | reason | — |
| vite | commit | coefficient | 0.452714 |
| vite | commit | pr auc delta | -0.015483 |
| vite | commit | baseline pr auc | 0.810068 |
| vite | commit | augmented pr auc | 0.794585 |
| vite | commit | eligible files | 79 |
| vite | commit | minor eligible files | 29 |
| vite | commit | prevalence | 0.696203 |
| vite | commit | minor count available | yes |
| vite | line | label | line ownership |
| vite | line | group | ownership source |
| vite | line | status | ok |
| vite | line | reason | — |
| vite | line | coefficient | -0.436127 |
| vite | line | pr auc delta | 0.008867 |
| vite | line | baseline pr auc | 0.810068 |
| vite | line | augmented pr auc | 0.818934 |
| vite | line | eligible files | 79 |
| vite | line | minor eligible files | 29 |
| vite | line | prevalence | 0.696203 |
| vite | line | minor count available | yes |
| vite | bots | label | include bots |
| vite | bots | group | data cleaning |
| vite | bots | status | ok |
| vite | bots | reason | — |
| vite | bots | coefficient | 0.534178 |
| vite | bots | pr auc delta | -0.015805 |
| vite | bots | baseline pr auc | 0.812231 |
| vite | bots | augmented pr auc | 0.796426 |
| vite | bots | eligible files | 80 |
| vite | bots | minor eligible files | 29 |
| vite | bots | prevalence | 0.700000 |
| vite | bots | minor count available | yes |
| vite | mailmap | label | without mailmap |
| vite | mailmap | group | data cleaning |
| vite | mailmap | status | ok |
| vite | mailmap | reason | — |
| vite | mailmap | coefficient | 0.452714 |
| vite | mailmap | pr auc delta | -0.015483 |
| vite | mailmap | baseline pr auc | 0.810068 |
| vite | mailmap | augmented pr auc | 0.794585 |
| vite | mailmap | eligible files | 79 |
| vite | mailmap | minor eligible files | 29 |
| vite | mailmap | prevalence | 0.696203 |
| vite | mailmap | minor count available | yes |
| vite | likely fix | label | likely-fix commits |
| vite | likely fix | group | outcome |
| vite | likely fix | status | ok |
| vite | likely fix | reason | — |
| vite | likely fix | coefficient | 0.452714 |
| vite | likely fix | pr auc delta | -0.015483 |
| vite | likely fix | baseline pr auc | 0.810068 |
| vite | likely fix | augmented pr auc | 0.794585 |
| vite | likely fix | eligible files | 79 |
| vite | likely fix | minor eligible files | 29 |
| vite | likely fix | prevalence | 0.696203 |
| vite | likely fix | minor count available | yes |
| vite | all change | label | all changes |
| vite | all change | group | outcome |
| vite | all change | status | ok |
| vite | all change | reason | — |
| vite | all change | coefficient | -0.051975 |
| vite | all change | pr auc delta | 0.001188 |
| vite | all change | baseline pr auc | 0.925557 |
| vite | all change | augmented pr auc | 0.926745 |
| vite | all change | eligible files | 79 |
| vite | all change | minor eligible files | 29 |
| vite | all change | prevalence | 0.835443 |
| vite | all change | minor count available | yes |
| vite | minor cutoff two percent | available | yes |
| vite | minor cutoff two percent | status | insufficient |
| vite | minor cutoff two percent | reason | training minority class 6 is below 8 fitted parameters |
| vite | minor cutoff five percent | available | yes |
| vite | minor cutoff five percent | status | insufficient |
| vite | minor cutoff five percent | reason | training minority class 6 is below 8 fitted parameters |
| vite | minor cutoff ten percent | available | yes |
| vite | minor cutoff ten percent | status | insufficient |
| vite | minor cutoff ten percent | reason | training minority class 6 is below 8 fitted parameters |
| vite | minor cutoff twenty percent | available | yes |
| vite | minor cutoff twenty percent | status | insufficient |
| vite | minor cutoff twenty percent | reason | training minority class 6 is below 8 fitted parameters |
| vite | minor cutoff fifty percent | available | yes |
| vite | minor cutoff fifty percent | status | insufficient |
| vite | minor cutoff fifty percent | reason | training minority class 6 is below 8 fitted parameters |
analysis-wide totals · 6 values
| repository | variant | claim | value |
|---|---|---|---|
| aggregate | aggregate | repositories | 5 |
| aggregate | aggregate | eligible files | 501 |
| aggregate | aggregate | minor eligible files | 89 |
| aggregate | aggregate | outcome files | 222 |
| aggregate | aggregate | primary models ok | 5 |
| aggregate | aggregate | minor models ok | 0 |
Contribution concentration is worth investigating when it changes a practical question. Here, it did not produce one stable result beyond size and activity.
References
- Bird et al., ESEC/FSE 2011. The original compiled-binary study and its observational failure association.
- Greiler, Herzig and Czerwonka, MSR 2015. A replication-oriented ownership study.
- Foucault et al., Information and Software Technology 2015. A study of ownership and quality measures.
- Borg, Tornhill and Mones, EASE 2023. Evidence on code ownership and maintenance work.
- Koana et al., Empirical Software Engineering 2024. A recent ownership analysis.
- Thongtanunam and Tantithamthavorn, 2024. An empirical comparison of commit- and line-based ownership measures and their associations with software quality.
- Naur, Microprocessing and Microprogramming 1985. The essay that describes programming as building a theory held by the people who write the code.
- Fritz et al., ICSE 2010. A degree-of-knowledge model that combines authorship with a developer's interaction with the code.
- Avelino et al., ICPC 2016. A related study of contribution concentration.