What do you want to achieve in the case? Get a decision changed, reach a settlement at a certain level or limit the financial risk?
Lestrea starts from your goal and the material in the case. The system explores possible strategies, tests how the other side may respond and helps you judge the next step.
The work rests on three parts that work together.
- 01Strategy Analysis tests every strategy.
- 02Probability Analysis shows how strong your position is.
- 03Strategy Paths shows your next move and the other side’s likely countermoves.
The analysis can then be used to build a draft with traceable support.
You get support to choose a path and keep working. You make the decisions.
The same case can have several possible goals. Winning in court is not always the same as reaching a quick agreement or keeping costs down.
That is why the work starts with the desired result. Appealing, for example, is an action. Getting the decision changed is the goal.
The goal gives the analysis a direction. The documents, the circumstances and the legal conditions decide which paths are relevant to explore. A legally possible move is not automatically the most useful one for what you want to achieve.
You add the documents and emails that belong to the case. Lestrea compiles the parties’ claims, key events, deadlines and evidence. The facts keep their link to the material they come from.
If two documents give different pictures of the same event, the difference can be examined. If an argument needs more support, it can become a question to investigate further.
The analysis separates what a source states, which conclusions are drawn and what is still an assumption. That something is stated in a document does not automatically mean it is proven.
Lestrea’s work also includes keeping track of which material has been examined and which questions remain. Missing or insufficient material should not be hidden behind a finished statement.
Which path is strongest?
A strategy can sound convincing and still rest on a weak assumption. An argument can have support but be irrelevant to the question to be decided. A next step can strengthen your position and at the same time open up a difficult objection.
The Strategy Analysis is the way of working that examines this. Proposals are produced, tested and reworked based on what the review shows.
Several approaches are explored
The first reasonable proposal is not necessarily the most useful one. Lestrea explores alternative strategies based on your goal.
What speaks for each path? Which material is needed? Which assumptions must hold? What do you risk losing or making harder by choosing that particular direction?
The comparison makes it clearer what a proposal rests on and which questions need answers before you move on.
Material, law and case law are tested together
Research is used to examine legal support and counter support. Lestrea works with legal rules and court cases together with the facts in the case.
The analysis also takes into account where the case is in the process, which court or authority is handling the matter and what it can actually decide.
It helps you tell the difference between an argument that sounds strong and an argument that is useful on the right question at the right stage.
The other side is tested against your own
In the adversarial review, also called red teaming, Lestrea examines how your strategy or your draft can be attacked.
Which fact can the other side challenge? Is there another possible interpretation of the material? Which counterargument would be hardest to answer?
The review does not only look for support for your position. It also looks for what can weaken it, so that you can prepare a response or change direction. The other side’s possible responses are treated as scenarios, not as known intentions.
The stress test examines what could make the strategy fall
A strategy can depend on a single decisive fact. If that fact is uncertain, it needs to be clear.
The stress test examines sensitive points: missing support, conflicting facts and objections that could change the assessment.
The question is not only what supports the strategy, but also what would have to be wrong for it to stop holding. The result shows what needs to be examined, strengthened or reconsidered.
The review affects the next proposal
When a weakness is found, it is fed back into the analysis. An argument may need to change, another approach may become more relevant or more material may be needed.
The analysis is not about repeating the same reasoning until it sounds certain. It should let objections and new information change the proposal.
Adversarial review
Lestrea tests your strategy from the other side’s perspective. The review examines which arguments can be attacked, which weaknesses need to be handled and which assumptions need to be checked.
This way of working is also called red teaming.
Examples of building blocks from the review
How strong is our position?
That a system states a 70 percent probability does not in itself mean the assessment is reliable.
Calibration examines the relationship between the probabilities stated and how often the corresponding outcomes actually occur. If an outcome is judged to have about a 70 percent probability in many comparable cases, it should occur in about 70 percent of them. If it only occurs in half, the assessments are too high.
The example explains the principle. It is not a measured result for Lestrea.
First, what the probability concerns is defined
The probability of winning in court is not the same as the probability of reaching a settlement. Having a particular claim granted is not automatically the same as reaching your whole goal either.
The assessment therefore needs to concern a clear outcome. When the forecast is followed up, the comparison must concern the same question.
A settlement cannot, for example, confirm an earlier forecast of winning in court, since that question was never decided by a court. Lestrea’s calibration model distinguishes between these outcomes.
The model starts from a statistical starting point
Lestrea’s calculation model uses a statistical starting point for the type of court and, where data exists, the type of case.
The model is Bayesian. This means that earlier knowledge works as a starting point that can be updated when relevant information is added. How much the assessment changes depends, among other things, on how extensive the material is.
The statistical starting point is not an answer key for your case. It needs to be kept apart from the circumstances and the evidence that are specific to your situation.
The same fact should not count as several independent sources of support
A fact can recur in several documents without having been confirmed by several independent sources.
The calculation model can reduce the combined weight of evidence signals that are connected. This prevents the same underlying fact from strengthening the assessment several times just because it is repeated.
So it is not only the number of facts that matters, but also how they relate to each other.
The forecast is linked to the outcome that later becomes known
To evaluate a probability assessment, the earlier forecast needs to be linked to a verified outcome in the right case and court context.
What did the model state before the outcome was known? What actually happened? Did the forecast and the outcome concern the same question?
Lestrea’s calibration flow is built for that link. The model’s own simulations may not replace real court outcomes when forecasts of success in court are evaluated.
Systematic deviations can be corrected
When relevant calibration data exists, the model can examine whether the probabilities are repeatedly too high or too low and adjust them.
This is different from changing the assessment after a new document in a single case. New information changes the basis for that particular case. Calibration examines how the probability assessments perform across many verified outcomes.
A changed percentage is therefore not automatically a better calibrated percentage. The improvement needs to be shown in an evaluation.
The evaluation decides which precision can be used
A calculated probability is not automatically a validated probability.
Lestrea therefore distinguishes between calculating forecasts, evaluating them and letting them influence strategic recommendations. Their use is subject to validation requirements, including calibration and the ability to tell different outcomes apart. The assessments need to hold in the relevant court context, not just in a pooled average.
When the material cannot carry a precise probability, the uncertainty should not be hidden behind an exact percentage. Broader assessments, ranges and direction may then be more relevant. A range is, however, not in itself proof of good calibration.
It should be clear what the number means
You need to be able to judge three things: which outcome the probability concerns, what material it rests on and how that type of forecast has been tested against real outcomes.
A comparison of strategies is not automatically a percentage for reaching your goal. The difference should be clear, so that you can weigh the assessment together with the legal material.
What is our next move?
The Strategy Analysis tests how sustainable a proposal is. The Strategy Paths explore possible continuations if you choose it.
A move is followed by a reaction that can change the conditions for the next decision. That is why an action needs to be judged as the start of several possible courses of events, not just as an isolated step.
Your move, the other side’s response and your next choice
A settlement offer can be accepted, rejected or met with a counteroffer. A request for more evidence can produce new material or leave the decisive question unanswered.
The Strategy Paths explore such branches. What can you do first? How may the other side react? Which alternatives may then arise?
It helps you prepare for more than the most favourable course of events.
Possible continuations are explored with tree search
The strategy engine uses a search model to explore possible moves and continuations. The alternatives can be described as a tree where every step opens new branches.
The purpose is to give a basis for the next choice, not to map the whole future. The number of possible courses of events grows quickly, and the uncertainty increases further ahead in the analysis.
What matters is which alternatives are worth considering now, what conditions they require and what could change the choice.
The paths are judged against your goal
If the goal is a settlement of SEK 200,000, the alternatives need to be assessed on that basis. A path that may give a higher amount after a long process is not necessarily better than an earlier agreement.
The difference depends on your priorities and which risks you are prepared to take. The goal sets the direction, but the path there needs to be tested against the conditions in the case.
When the other side responds or new material is added, the alternatives can be explored again.
Once you have chosen a direction, the parts of the analysis can be put together into a draft. You choose which arguments to use, which material should support them and which objections need to be answered.
Blocks can be added, removed, replaced or reviewed again. The text that connects them is also covered by the document’s source checks.
The source links follow into the draft. You can go from a claim to the relevant material and check that the wording has support. The draft is working material that you review and edit, not an automatically approved document.
Review in Word
Drafts can be exported to Word as review material, with or without tracked changes. When changes are brought back into Lestrea, they are first shown as review suggestions.
Accepting a change in Word is not the same as approving its legal content. Accepted changes need to be brought into the source bound working material and checked before the document can be approved in the rest of the flow.
Lestrea helps you explore the alternatives, discover the weaknesses and understand what the assessments rest on. You choose the direction and take a position on the material.

