Tool

Your recall is your ceiling.

If the passage carrying the answer never gets retrieved, nothing downstream recovers it. This works out the cap that puts on your system, splits your errors into the two halves that need different fixes, and tells you which one your budget belongs in.

Your two measurements

Take fifty questions your system gets wrong. For each, check whether the passage containing the answer appeared anywhere in what was retrieved. That single split, one afternoon of work, gives you both numbers below and settles where the budget belongs.

From real users, with correct sources set by someone who knows the domain.
72 %
Share of questions where the answering passage was retrieved at all.
88 %
Same measurement at a larger candidate set. The gap is what reranking can reach.
86 %
Given the right context, how often is the answer right?
12 %
Right anyway, from what the model already knew. Usually low, and worth measuring rather than assuming.
This calculator runs entirely in your browser. Nothing you type is sent anywhere unless you ask for the result by email at the bottom of the page.
The ceiling
ScenarioEnd to endRightWrongWhat it tells you
Today65.3%326174the number you have
If generation were perfect75.4%377123the cap recall puts on you
If retrieval were perfect86.0%43070the cap generation puts on you
Reranking the wider set77.1%386114reachable without changing the index
Counts are out of your evaluation set, so they are comparable against a hand-check.
End-to-end accuracy
65.3%
Ceiling set by recall
75.4%
A perfect model cannot pass this
Errors caused by retrieval
71%
The larger half
Gain available from retrieval
+20.7%
Gain available from generation
+10.1%
Available from reranking alone
+11.8%
No index change, no new embeddings
Retrieval is the constraint. Even a perfect model reaches only 75.4%, because 28% of questions never see the passage that answers them. A good share of that is reachable by reranking a wider candidate set, which is far cheaper than re-embedding the corpus.
Send me this split

Your two numbers go with it. If retrieval is the larger half and the budget is going to the model, that is usually the whole finding.

Your inputs are included so the reply can be specific.

End-to-end accuracy, by recall and grounded accuracy

What a RAG system delivers overall, given how often it retrieves the right passage and how often it answers correctly once it has. Read across a row to see how little a better model buys when recall is low. The final column is a perfect model, which is the ceiling recall alone imposes.

Recall70% grounded80% grounded90% grounded95% groundedPerfect model
50%40.0%45.0%50.0%52.5%55.0%
60%46.0%52.0%58.0%61.0%64.0%
70%52.0%59.0%66.0%69.5%73.0%
80%58.0%66.0%74.0%78.0%82.0%
90%64.0%73.0%82.0%86.5%91.0%
95%67.0%76.5%86.0%90.8%95.5%

Assumes the model answers correctly without the passage 10 percent of the time, which is a typical figure worth measuring rather than inheriting. Red is under 70 percent, amber under 85. Note the last column: at 70 percent recall a flawless model still delivers 73 percent.

A worked example

A support assistant over an internal policy corpus, reported at 65 percent end-to-end. The team's plan for the quarter is a better embedding model and a prompt rewrite, on the reasoning that the answers read as though the model is not understanding the material.

Fifty failed questions get split by hand in an afternoon. In 36 of them the passage containing the answer was never retrieved. Recall measures 72 percent; when the passage is present the answer is right 86 percent of the time. That puts end-to-end at 65 percent, which matches the reported number, and it means 71 percent of the errors are retrieval.

The ceiling is the part that changes the plan. Even a flawless model, on this index, reaches 75 percent. The entire prompt-and-model quarter is competing for at most ten points, and most of it is not available at all. Meanwhile recall at a wider candidate set is 88 percent, so a reranker over passages already being retrieved is worth about twelve points on its own, with no re-embedding and no migration.

The order that follows is recall first, reranking second, generation last. It is the opposite of the plan, and the afternoon that produced it cost less than the first week of the quarter would have.

The arithmetic, so you can check it

End-to-end accuracy is recall x grounded accuracy + (1 - recall) x accuracy without the passage. Set grounded accuracy to 1 and you have the ceiling recall imposes; set recall to 1 and you have the ceiling generation imposes. Neither can be exceeded by working on the other side.

The error split is the same expression rearranged: retrieval errors are (1 - recall) x (1 - accuracy without the passage) and generation errors are recall x (1 - grounded accuracy). Reranking headroom is the recall gained by widening k, multiplied by the accuracy difference that passage makes.

The honest limit

This treats retrieval as binary: the passage is in the set or it is not. Real systems have gradations that matter, including a passage retrieved but ranked so low it is truncated out of the context, and answers that need two passages where only one arrived. Both make your effective recall lower than the measured figure, so treat this as an optimistic bound. It is still the right bound to argue from, because it is already enough to settle which half the budget belongs in.

The reasoning is in recall is the silent killer and the split procedure in retrieval or generation, which half broke.

Questions

What is retrieval recall and why does it cap everything?

Recall is the share of questions where the passage containing the answer made it into the retrieved set at all. If it is not there, the model is being asked to answer from material that does not contain the answer, so no prompt, no larger model and no reranker recovers it. End-to-end accuracy cannot meaningfully exceed recall, which makes recall the ceiling on every downstream improvement.

How do I measure recall without a labelled dataset?

Take fifty questions from real users that your system gets wrong, and for each one check whether the passage containing the answer appeared anywhere in what was retrieved. That is an afternoon of work and it splits your failures into the two buckets that need completely different fixes. Correct sources should be set by someone with the domain knowledge to judge them, not by the engineer and not by a model.

Should I improve retrieval or generation first?

Whichever half holds more of your errors, which this computes. The common mistake is spending a quarter on prompting and model choice when most failures are questions where the passage was never retrieved. Improving the half that was already working produces no visible change and is the most expensive way to gather no information.

When does reranking help?

Only when recall rises appreciably as you widen the candidate set. That gap is the material reranking has to work with: passages already being retrieved but ranked too low. If recall is flat and low across k, the passages are not being found at all and a reranker has nothing to reorder.

Why does the calculator ask how often the model is right without the passage?

Because it usually is not zero, and assuming zero understates your ceiling. Some questions the model answers correctly from what it already knew. That rate is worth measuring rather than guessing, and it is also the rate at which your system produces right answers for reasons your retrieval cannot explain or attribute.