Alliance Selection
Ranks the teams you have scouted by what they are actually likely to contribute, and says which gaps between them your data can genuinely support. Four matches of a streaky robot proves less than you think.
Inputs
One row per team you would consider. The spread is how much they vary between matches: take the highest and lowest scores you saw and divide the gap by four for a rough figure, or work it out properly if your scouting sheet has every match.
Your own average, on the same basis. Only used to show the projected alliance total, not the ranking.
How sure you want to be that one team really is better than the next before treating the gap as real. 80% is a reasonable bar for a pick you have to make in ten minutes.
Results
Best available
—
Their average
—pts
Their true average is near
—pts
95% interval on the average, not on a single match.
Confidence they beat the next
—%
Is the gap real?
—
Projected alliance total
—pts
The list, ranked
Ordered by average. The interval is on the average itself, so it narrows as you watch more matches even if the team stays just as streaky.
| Team | Matches | Average | Spread | True average within | Beats the next |
|---|
How this is calculated
An average from four matches and an average from eleven are not the same kind of number, even when they read the same. How much you can trust one depends on how streaky the team is and how many times you saw them, and those two things combine into a single figure.
standard error = spread ÷ √matches interval = average ± 1.96 × standard error confidence A > B = Φ( (avgA − avgB) ÷ √(seA² + seB²) )With your numbers
The standard error is the spread of the average, not of the team. A robot that swings wildly can still have a well known average if you watched it enough times, and a consistent robot seen twice can still surprise you.
Why the highest average is not automatically the pick
A team averaging 46 over four matches with a big spread might genuinely be worse than one averaging 41 over nine with a small one. The confidence column is the honest answer to that: below the bar you set, the ranking is not something your data supports and you should pick on something else.
That something else is usually what they do rather than how much: whether they can play the part of the strategy you cannot, whether they break, and whether they can be told where to go. None of that is in this table.
What this does not know
Points scored with a strong partner are not evidence of what a team does alone, and scouting data rarely separates the two. It also assumes the matches you watched are representative, which stops being true the moment a team fixes something halfway through the day. If they had a rebuild, only count the matches after it.
Sources & assumptions
Everything on this page is yours. There is no scouting database here and there never will be: numbers for teams at your event come from watching them at your event.
The interval is a normal approximation on the mean, which is standard and reasonable once you have a handful of matches. The confidence that one team beats another compares two means with their standard errors. Both are derived, with the formulas above.
- V5RC Override Game Manual v1.1 (2026-27) — checked 2026-08-17
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.