iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A commit count tells you how many times a person saved a snapshot of their changes to Git. It does not tell you how much engineering value those changes carried. In a personal essay published on DEV Community, software engineer László Szabó describes being told his individual commit count trailed some coworkers’, then building an agent skill called crazy-commiting that split his existing changes into more, smaller commits. In his account, the dashboard number rose a few weeks later even though the feature, the code and the effort were the same. His point is that the metric measured how work was recorded, not what the work delivered.
What happened, in the author’s account
Szabó says his employer tracked individual commit counts as an engineering performance metric and told him his number was lower than some colleagues’. He describes his role at the time as covering architecture, technical decision-making, mentoring, code review, team leadership, difficult debugging, cross-product coordination and coding. By his count, many of those responsibilities produced no commits of his own. His blog post is dated 28 September 2026 and places the reprimand in 2025. The essay does not describe the company’s KPI system beyond that, and the details are his own.
How the skill worked
Szabó built crazy-commiting as a post-work agent skill. He ran it after he had finished and reviewed a piece of work. According to the essay, the skill was meant to:
Free tools Windows power users keep installed
One-click scans. No signup required.
- Inspect the pending changes in the working tree.
- Identify parts of the change that could stand on their own as coherent units.
- Stage each part separately.
- Write a proper commit message for each one.
The stated target was the maximum number of reasonable, coherent commits. Szabó says it was explicitly not meant to produce fake commits, whitespace-only changes or empty messages. His example turns one broad synchronization commit into separate commits for configuration, repository access, mapping, service logic, validation, error handling and tests. Each of those commits is a defensible unit in its own right, which is exactly why the count rises without any new work being done.
#1 Best Overall
Why one commit is not a unit of value
A commit has no fixed size. A typo fix and a complex data migration can each be one commit. The same final change can also be recorded as one commit, four or seventeen, and the repository ends up in the same state either way. The table below shows the point using the essay’s illustrative numbers.
| How the same change is recorded | Commits counted | Final repository state |
|---|---|---|
| One broad commit | 1 | Same as the other rows |
| Split into logical units (configuration, access, mapping, logic, validation, errors, tests) | 4 to 7, depending on grouping | Same as the other rows |
| Split into fine-grained commits | 17 | Same as the other rows |
Where commits are counted matters too
The author’s blog adds a measurement caveat. Depending on where a dashboard counts, squash-merging can make a branch with many commits appear as a single commit on the main branch. A count taken on the main branch and a count taken on feature branches can therefore disagree about the same work, and neither is a measure of effort.
What activity counts miss
For senior and lead roles, Szabó lists work that rarely shows up in a commit log: code reviews, mentoring, system design, production incident investigation, migration coordination, risk reduction and helping the team avoid unnecessary complexity. He makes the last point concrete. A decision not to build a service nobody needs can produce no lines, no commits and no pull requests, yet it may save months of maintenance. These are his reasons, not measured results, and the essay does not quantify them.
- Reviews and mentoring usually leave no authored code behind.
- Design and incident work often happens before or around the code, not in it.
- Preventing unnecessary code is invisible to any count of additions.
- Coordination across products or teams is recorded in meetings and tickets, not in commits.
Where activity data still earns its place
Szabó does not argue that activity data is worthless. An unusual change in repository activity can be useful context and a reason to ask a question. The error he describes is skipping that context and turning the graph into a conclusion. His summary is that “the commit graph can help start the conversation.” The practical question he suggests is simple: when activity drops, ask “What are you working on?”
Role-appropriate expectations and outcome-based goals
The author’s blog post recommends setting goals around outcomes and judging people against expectations that fit their role. Examples of outcomes include a migration shipping, an incident rate falling, a new hire becoming productive, or an architecture decision holding up under load. He suggests that leads be assessed on team delivery, technical decisions and the growth of the people they work with, while individual contributors can be measured more closely on their own output.
DORA
The DEV article points readers to DORA, which it describes as focusing on how software moves through an organization. The author’s blog names four DORA metrics:
Rank #3
- Deployment frequency
- Lead time for changes
- Change failure rate
- Time to restore service, which the blog says has recently been renamed failed deployment recovery time
DORA’s official site identifies it as a Google Cloud program studying the capabilities that drive software delivery and operations performance.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallSPACE
The author summarizes SPACE as five dimensions: satisfaction and well-being; performance; activity; communication and collaboration; and efficiency and flow. This is a secondhand summary. The original SPACE article could not be checked for this piece, so anyone using the framework in a performance review should read the primary publication for its exact definitions.
Further reading
The author’s blog names Accelerate: The Science of Lean Software and DevOps in its discussion of software delivery performance. It is a reasonable next read for anyone who wants the background behind DORA, but the essay does not endorse a particular edition or retailer.
Rank #4
What AI agents change about the metric
Szabó’s broader argument is that AI agents make visible activity cheap to produce. Commits, pull requests, lines of code, tests, documentation and tickets can all be generated faster than the work they describe. In his example the agent wrote no additional code. It changed how existing changes were represented in Git history. This is his prediction and interpretation, not a measured industry-wide finding.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Goodhart’s Law and the question to ask
The essay frames the problem with Goodhart’s Law, in the form “When a measure becomes a target, it stops being a good measure.” The essay does not identify an original source for that exact wording, so it should be attributed to the article if quoted directly. Szabó’s closing question is the one a manager should ask before acting on any activity chart: “If I can improve the metric significantly with an agent without improving the product, the team, or the engineering outcome, what exactly is the metric measuring?”
For a manager, the useful reading of a commit graph is as a prompt to talk with the engineer, not as a score. A drop can mean a stalled task, a review-heavy week or a design decision that removed work. A rise can mean better commit hygiene. The graph cannot tell those apart, and the conversation is how you find out which one you are looking at.
Best Value
- Author: Bungay Stanier, Michael.
- Publisher: Page Two
- Pages: 244
- Publication Date: 2016-02-29
- Edition: 1
What the evidence does and does not establish
- The essay is a personal account. It is not a controlled study and not an independent investigation of the employer.
- The company’s KPI system, the reprimand and the dashboard change are reported by the author and have not been independently verified.
- The essay gives no measured effect size for the skill and no organization-level productivity result.
- The DEV Community page shows “Posted on Sep 29” without a year. The linked original blog post is dated 28 September 2026.
- No public repository for
crazy-commitingwas identified, so the skill cannot be inspected from the essay alone.
Within those limits, the essay makes a narrow but sound point. Commit counts reward how changes are split and where they are counted. A team that wants to know whether engineering is healthy needs outcomes, and a manager who wants to know about a person needs to ask them.
Each sentence reflects the author’s account and his interpretation. Readers should weigh the reasoning on its merits rather than treating the numbers in the example as evidence.
Frequently asked
This article answers the questions a reader is most likely to ask, so there is no separate FAQ section.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsQuick Recap
“
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
