Technical reports aimed at estimating compensation for the frustration of profits in a construction project, which one party ceased to receive as a result of damage caused by the other party, often rely on the presumption of a counterfactual scenario of how events would have unfolded in the event that the damaging incident had not occurred. This is done with the aim of accrediting and proving the causal link between the damage and the frustrated profit. Estimating compensation for frustrated profit presents the challenge of requiring rigorous evidence to establish its exact scope and proof. Arbitration tribunals may not only require a projection of the profits foregone but may also demand an objective possibility of success for the project in question, considering variables that are not based on vague and uncertain assumptions. Therefore, the expert report, in addition to having an economic and financial methodology, must also substantiate the hypothesis of the project’s success probability in order to build a consistent valuation model.
In this sense, the “ANETRIC RNA Engineering” methodology proposes a predictive model of the probability of success or failure of a construction project before it begins. Assuming the existence of a project with a signed contract, whose commencement of execution was frustrated by an action attributable to the contracting party. The predictive model is based on a risk assessment mechanism for a project, through the creation and use of a database with historical indicators to measure the impact of risk factors on the objectives of an infrastructure project, under a profitability scheme. This involves comparing historical data with risks identified in the business, using artificial neural networks as an analytical tool.
However, predicting the success or failure of a project involves predicting the risk of success or failure in its execution, so it is necessary to assess the risk that could be assumed by choosing an appropriate metric. Risk should not only be evaluated in terms of the probability of project success or failure, but also in terms of the importance of the goal to be achieved. Therefore, it is important to propose measurement indicators for risk factors that impact project objectives under a profitability scheme, applying a control mechanism that compares current risks with historical data from past projects of the organization, using artificial neural networks as an analytical tool.
It is necessary to establish a model that estimates with the least possible error the probability of success or failure of an infrastructure project under normal and foreseeable conditions, before it is initiated. This is because we assume that the agreed project was frustrated before it started due to an action attributable to the contracting party. Therefore, the expert report, in addition to having an economic and financial methodology to quantify frustrated utility, must also substantiate, in the construction of its hypothesis, the probability of project success in a solid manner. This is because if the probability is that of failure, positive financial projections would not make sense.
However, it is possible to subjectively calculate the probability of success or failure of a project before it starts, based on consultations with experts or direct consultations with all parties interested in the project. This should provide a subjective rating for each identified risk. However, this risk estimation process contains too much uncertainty and cannot be considered reliable. Similarly, it is possible to make an objective estimation of the project’s risk before it starts based on an evaluation of variables by comparing threshold values with similar projects executed by the affected contractor. This can quantitatively classify the project before it starts as a success or failure. However, the problem that may arise is that the threshold values considered may contain too much uncertainty to define the comparison threshold, as it is simply based on a weighted average of exposure to risk from historical data of the organization, or in some cases, the problem may arise where the company adopts theoretical comparison values (for example, values belonging to other organizations) to predict the success or failure of a project. In both cases, the risk estimation process contains too much uncertainty and is not reliable.
Solution Strategy:
The proposal is to establish a model that uses historical data from similar projects executed by the organization to identify risk factors affecting project success variables. The proposed model uses artificial neural networks as an analytical tool to determine the comparison thresholds for success or failure of a project before it starts. The proposed model will be built based on historical data from similar projects executed by the organization, using artificial neural networks as a tool, and will have as its output or estimated evaluation variable the project’s objective indicator (effectiveness, efficiency, and quality), and risk calculation in terms of the probability of success or failure of the project before it starts. Therefore, the challenge is to determine to what extent risks can impact project objectives. For example, if risks have a significant impact on the project, we should expect a low probability of success. However, we must take into account the basis or threshold of comparison that the model provides. For example, suppose the project’s objective is effectiveness, measured by the effectiveness index (results achieved over planned results) to define the success or failure of a project, and we compare it with the effectiveness indices of similar projects in the company. Suppose for a project, the Effectiveness index “E1” is between 0.7 and 0.8. If a theoretical effectiveness index of success E2=1.0 is defined as the comparison threshold, then the project is considered a failure because E1E3). This occurs because the comparison basis is based on real experiences of the organization. Therefore, we must cal
ibrate the evaluation criteria of the observed data based on the actual performance of the company, ensuring that the information is sufficient and as up-to-date as possible to estimate an optimal model that uses artificial neural networks as a tool. The proposed strategy is based on designing common historical indicators for projects of an organization, allowing the construction of a historical database of the organization’s behavior in the execution of similar projects, in order to calculate the impact of risk factors on project objectives. The purpose of this strategy is to generate a comparison basis for the value of risk.
