Intake Jam Reliability Analyzer
Count what actually happened over a run of attempts and find out which version is better, and whether you have enough trials to say so.
Inputs
One row per version you tested. The four outcome columns should add up to the attempts, and the tool says so if they do not.
Results
Best version
—
Its success rate
—%
95% confidence range
—%
Beats the runner-up?
—
Attempts recorded
—
Most common failure overall
—
Version by version
| Version | n | Success | 95% range | Jams | Doubles | Bounce | Worst failure |
|---|
How this is calculated
A success rate on its own is close to meaningless. Twenty-one out of thirty is 70%, but run the same intake again and you might get 24, or 18. The question is not what happened, it is what the rate would settle at if you kept going.
Why the range matters more than the rate
The 95% range is a Wilson score interval, which asks what true success rates could plausibly have produced the result you saw. Its width is set almost entirely by how many attempts you ran.
The practical consequence is blunt. Two versions whose ranges overlap have not been told apart. Version B looking better than version A on twenty trials each is usually noise, and rebuilding around that difference is rebuilding around nothing. Run more trials, or accept that you cannot yet tell.
The interval also handles the case a simpler formula gets wrong. Ten successes out of ten does not mean a perfect intake; it means the true rate is probably somewhere above 70%, which is a very different claim.
With your numbers
Separate the failures
Lumping every failure together hides the fix, because the three modes pull in opposite directions:
- Jams usually mean too much compression, or a gap narrower than the object.
- Double feeds usually mean too little compression, or no feature to separate objects.
- Bounce-outs usually mean too much speed at the mouth, or nothing to trap the object once it is in.
Note that the first two have opposite remedies. An intake with both is being asked to handle objects that vary more than the mechanism tolerates, and tightening the compression will trade one failure for the other rather than removing either.
Running the test honestly
Vary one thing at a time, keep everything else identical, and use the same objects in the same condition. Alternate between versions instead of doing all of A then all of B, so a drifting battery or a warming motor does not land entirely on one of them.
Sources & assumptions
No VEX data is used. The statistics are applied to counts you record, so nothing here goes stale between seasons.
The interval is a 95% Wilson score interval. It assumes attempts are independent and identically conditioned, which is only true if you kept the test consistent. It cannot detect a version that works on a fresh battery and fails on a flat one.
Save this run, and compare
Keeps what is on screen so you can change something and see both sides of the change. Saved in this browser only, never uploaded.
Save this as evidence
Collects what you entered, what came out, how it was worked out, and anything the tool flagged, with a timestamp and a version so someone else can reproduce it.
This is evidence, not a notebook entry. It deliberately does not write your problem statement, your reasoning, or your conclusion, because under RECF rules an Engineering Notebook has to be the students' own work and no tool may generate or organise its content. Take the numbers, decide what matters, and write it yourself.