FindR is a data-driven technology platform developed to evaluate, classify, and prioritize identified targets in oil, natural gas, mining, and geothermal projects, as well as to support decision-making processes. FindR’s fundamental innovation does not rely on a single sensor, a single physical measurement, or a single geophysical parameter. The system’s core structure is built on the combined use of a multi-layered and comprehensive reference database with proprietary data processing, similarity analysis, and pattern-matching algorithms.
The fundamental concept underlying the method is as follows: Geological environments where hydrocarbons or mineralization were found in the past, as well as areas where economic or expected results were not achieved in past field studies, share specific combinations of measurable geological, physical, chemical, surface, soil, and drilling-related characteristics. FindR uses these past sites with known outcomes as a reference population and aims to assess the extent to which the multidimensional characteristics of a new exploration area resemble conditions associated with past successful or unsuccessful outcomes.
Therefore, the system’s objective is not to make a direct and definitive claim about a source based on a single isolated signal; rather, it is to generate a relative target priority by comparing the multivariate characteristics of the new area with those of sites where results have been verified in the past.
The fundamental approach of the FindR reference data bank is to systematically bring together the locations, well/borehole information and final field results of water, mineral, oil, natural gas and geothermal drilling operations carried out in the past.
The core principle underlying the data bank is the use, as far as possible, of data that are measurable, verifiable, capable of being placed on record and confirmable by field or laboratory results.
Data verified directly through drilling, laboratory analysis or measurable field results in particular constitute the core and strongest reference group of the FindR data bank. The aim of this approach is to establish a database that is as objective, comparable and re-verifiable as possible in the evaluation of a new exploration area.
In the FindR data bank, "measurement results" and "interpretation results" are deliberately separated from one another. Directly verifiable information — such as the actual lithology encountered during drilling, laboratory values, the presence or absence of fluid, temperature, flow rate, assay results and the final well result — is included in the principal reference data group.
By contrast, interpretation products derived from geophysical technologies and open to expert interpretation, modelling preference or differing technical assessments are not included in the principal learning and matching data group of the FindR reference data bank. Similarly, data obtained from remote sensing applications are likewise not taken into account within the scope of FindR's principal evaluation and matching data group.
This choice does not represent an approach that questions the value or the fields of application of geophysical technologies or remote sensing applications. Seismic, gravity, magnetic, MT, resistivity, IP, remote sensing and other conventional methods are important elements of natural resource projects in line with their own technical purposes. The data selection of the FindR data bank, however, is methodologically intended to rest as far as possible on result data that are directly measurable, verifiable and confirmable by field or laboratory results.
The fundamental question of the data bank is this: which results were obtained at which sites in the past, and under which measurable geological, physical, chemical and field conditions did those results occur? To what extent does the new site resemble these verified past examples?
For this reason, the reference point of the FindR algorithm is, to the greatest extent possible, actual drilling results, actual laboratory data and direct field observation. Holding successful and unsuccessful results together is intended to enable the system to learn not only from positive examples, but also from the conditions that yield no economic result.
At a new exploration site, representative samples and the available measurable characteristics of the site are used wherever possible. The sample is analysed in the laboratory to determine its physical and chemical properties. Depending on the nature of the project, geological characteristics, surface and soil parameters, chemical composition and other measurable variables meaningful in terms of the target resource may also be included in the assessment.
These data are standardized and converted into a multidimensional feature profile. In other words, the new site is represented not by a single value, but by a characteristic data profile formed by a large number of interrelated measurable variables.
The importance of the sample lies in its provision of direct data on the actual and current physical–chemical condition of the new site. The algorithm compares this current profile with past reference areas gathered from around the world within the data bank.
Where geochemical sampling has previously been carried out at the partner company's operating sites, FindR requests, during the evaluation process, the coordinates of the locations from which the samples were taken and the associated geochemical analyses/reports. Insofar as they are available, these data are included in the assessment within the scope of verifiable measurement and laboratory data relating to the site.
FindR's proprietary algorithm compares the multidimensional characteristics of the new area with past site populations in the data bank; it identifies the most relevant reference groups, examines the past results associated with these groups, and produces a relative target priority assessment.
The computation and evaluation process may include the following operations:
This process is not a simple database search. The aim is to identify not merely the new area's similarity to a single past site, but the most meaningful comparison populations, in geological and physical–chemical terms, from among a large number of past observations.
Depending on the size and heterogeneity of the data set, the data preparation required and the analytical scope, the evaluation period for a project may range from several weeks to several months. This period depends on the scope of the data preparation, quality control, standardization, comparison and iterative evaluation processes.
The details of FindR's algorithms, data processing methods, application procedures and operating principle are not shared publicly, as they constitute the company's intellectual property and its know-how in the nature of a trade secret.
The publicly available technical description is limited to explaining the general operating framework of the technology, the data groups used, the evaluation approach and the verification logic. Algorithmic weightings, spatial processing details, data transformation methods and other company-specific computational steps are protected as trade secrets.
What happened in the past? Under which geological, physical and chemical conditions did it occur? Where are the same or similar conditions found today? Which results were obtained in the past in areas possessing these conditions?
The FindR algorithm seeks the answers to these questions within its extensive reference data bank and thereby endeavours to produce a relative potential and target priority for the new exploration area.
In oil exploration, FindR uses as reference the physical, chemical, geological, surface and other measurable characteristics of areas where hydrocarbons were discovered in the past, where no hydrocarbons were found, or where drilling was completed with predominantly water-bearing or uneconomic results.
The algorithm seeks to produce answers to two complementary questions:
This assessment produces not a geological guarantee, but a relative probability or target priority. High-priority areas are subsequently tested in detail with conventional geology, geophysics, well logs, pressure, fluid, drilling and production data.
In mining, the same framework may be applied to gold, copper, lithium, nickel and other targets. The geological, geochemical, soil, surface, physical and laboratory characteristics of the new area are compared both with areas where mineralization was confirmed in the past and with areas where economic mineralization was not confirmed in drilling or sample results.
The system seeks to identify zones whose combined characteristics more closely resemble environments that were mineralized in the past, and to classify and prioritize targets within a large licence area on a relative basis.
Similarity alone is not evidence of an economic deposit. The final determination of resources and reserves requires drilling, representative sampling, laboratory assays, geological and structural interpretation, three-dimensional modelling, continuity assessment, metallurgical testing where necessary, and independent technical evaluation.
The same fundamental data-matching approach may also be applied in geothermal exploration. Past geothermal field and well data, the geological setting, physical properties and physical–chemical characteristics may serve as a reference in the evaluation of a new area. The aim is to narrow the broad exploration site and direct advanced geophysical and drilling work towards priority targets.
The conventional exploration group assumes the fundamental role in identifying potential reservoirs through geology, geophysics, seismic and other methods. FindR does not replace this process; it provides complementary decision support at the stage of classifying the reservoir targets identified by the company in terms of their content and producible hydrocarbon potential, and of determining drilling priority.
FindR's fundamental difference is that, rather than evaluating a new site solely through a theoretical geological model, it compares that site against a broad data population consisting of real-world drilling operations whose results are known and of field analyses carried out in the past.
The system's core asset is the quality and scale of the reference data; its core technological element is the proprietary algorithm that cleans, standardizes, compares, weights and interprets these data.
To date, various high match rates have been reported in FindR's internal field records. For the 79 boreholes in oil exploration whose results have been completed, the algorithm outputs are stated to be consistent with the field results. It is further noted that, in certain data sets and cases tested, target match rates of up to 99.99% have been reported.
REAL FIELD TESTING IN OIL, AT 124 DRILLING LOCATIONS
FindR's claim does not rest solely on theory or laboratory work.
To date, 124 oil drilling locations determined at the conclusion of conventional exploration processes have been accessed independently, and for each location only two basic predictions have been produced:
"There is producible oil or gas." "There is no producible oil or gas."
A portion of these locations are the points at which companies decided to drill in their fields, sent to us by those companies; the remainder consists of well locations and well names obtained by FindR for the purpose of independent testing.
The data for these wells were obtained independently through local sources of information in the areas near the drilling points and through third parties working in the drilling operations, rather than directly from company executives or decision-makers.
Once the well locations and names were reported to FindR, the predictions were produced, placed on record, and subsequently compared with the field drilling results received from the persons through whom we obtained access to the well information.
To date, drilling has been completed at 79 of the 124 wells.
RESULTS AT 79 WELLS
Thus, at all 79 wells for which results have been obtained, the prediction matched the field result.
Two of these 79 wells are offshore oil and natural gas drilling locations.
The field results for the remaining 45 wells are being monitored, as they have not yet been completed. As results are obtained, FindR's overall performance will be reassessed.
One of the most striking examples demonstrating FindR's precision is that of two wells in the same field, only 164 metres apart.
For the first well, FindR predicted:
«"Drilling is being carried out at the correct point, over the selected oil reservoir area, and oil will be found here."»
When the drilling was completed, the result matched FindR's prediction entirely, and oil was found in the well.
However, for the second well, just 164 metres away, FindR produced a different result:
«"This drilling is being carried out outside the oil reservoir area. It will end in failure and will be a dry hole."»
The drilling result once again coincided entirely with FindR's prediction.
This example shows that, even though the two wells are only 164 metres apart, FindR was able to distinguish between a full oil reservoir area and an empty area, or one containing no producible oil, and between the correct drilling point and the incorrect drilling point.
Likewise, in a field with more than 70 production wells, FindR correctly predicted the outcome of 5 dry holes weeks and months in advance.
In another field, the locations and well names of three neighbouring wells were reported to FindR, and a prediction was requested regarding the drilling results.
FindR's assessment was:
«"Yes, these three boreholes are drilled over a reservoir area; however, there is no producible oil here."»
When the drilling was completed, oil shows came from the wells. The field team was thereupon obliged to decide whether the wells would be closed as dry holes or brought into production.
In the end, it was decided to bring the wells into production. However, FindR maintained its earlier prediction and stated:
«"There is no producible oil."»
After the wells were brought into production, they produced only 401 barrels of oil in total and were subsequently recorded as dry holes on account of the water–oil balance. Production was halted and the wells were later converted into injection wells.
The significance of this example is that it shows the question is not, in fact, whether a reservoir is present or not. The reservoir is there, and oil shows are present as well. The critical question, however, is whether the reservoir contains producible oil in quantities capable of delivering economic and sustainable production.
FindR's prediction was realized at precisely this point:
The reservoir was there, but there was no producible oil.
Another company shared with FindR the drilling location at which it believed oil to be present at the conclusion of conventional exploration processes.
According to FindR's assessment, the selected location was indeed within a reservoir area, and there was oil within the reservoir as well. In other words, the choice of reservoir was fundamentally correct.
However, there was a critical problem:
The well location was very close to the outer boundary of the oil reservoir area.
For this reason, FindR predicted that oil production could be carried out at the well for a time, but that production would not be sustainable and that, shortly afterwards, the well would cease to produce and be recorded as a dry hole.
The field result again coincided with FindR's prediction. The company carried out production at the well for a period but was subsequently obliged to shut it in.
This example shows that selecting the correct reservoir is not sufficient on its own; selecting the correct drilling point within the reservoir is also critical to drilling success.
At one of the 124 drilling locations, when we met the owner of an oil company, the company had made a new discovery in its fields and had brought a well into production.
The company official shared with FindR the location of a new well of which only the first 100 metres had been drilled and for which a total depth of 2,700 metres was targeted, and asked for our prediction regarding the drilling result. The distance between the two wells was 700 metres.
In order to carry out our assessment, we also requested an oil sample — about the size of a small water bottle — from the well already in production in the same field, and it was sent to us.
As a result of FindR's assessment, the prediction that the new well was positioned at the correct point and over the oil reservoir area, and that "producible oil would be found" in the well, was communicated to the company.
The notable aspect of this work was that the assessment was not limited solely to the prediction of "producible oil present."
There was an analogue well in the same field that had been brought into production and whose production volume was known. Comparing the data obtained from this well with the assessment relating to the new well, FindR predicted that the new well's daily oil production could be approximately 60–70% higher than that of the analogue well.
About a month and a half passed. As the new well approached the target depth, according to the field information shared by the company official, a 21-metre oil "pay zone" was intersected in the well. In the analogue well brought into production in the same field, this figure was around 10 metres.
These field findings appear consistent both with FindR's prediction of "producible oil present" and with its assessment that the new well would have a higher production potential than the analogue well.
However, as the well has not yet been brought into production, the prediction regarding production volume is not yet a finalized result. When the well is brought into production, the assessment will be compared once again against the actual production data obtained.
This example demonstrates that FindR's aim is not only to answer the question "is there oil or not?" but also, when evaluated together with available analogue data, to produce predictions relating to production potential.
THE FUNDAMENTAL DIFFERENCE FINDR PRESENTS
The examples FindR has obtained to date reveal three important distinctions:
The distinction between a full and an empty reservoir.
The assessment of whether a reservoir area genuinely contains producible oil.
The distinction between oil shows and producible oil.
Demonstrating that the presence of oil shows in a well does not mean that the well will deliver economic and sustainable production.
Assessment of the reservoir boundary and the drilling point
In addition to selecting the correct reservoir, assessing where the drilling point lies within the reservoir and how close it is to the boundary.
In other words, FindR's aim is not merely to answer the question:
"Is there a reservoir?"
but to produce a prediction addressing the question:
"Does this reservoir contain producible hydrocarbons, and is this the right target for drilling?"
FINDR'S PLACE IN THE CONVENTIONAL EXPLORATION PROCESS
As a result of geological and geophysical studies spanning years, companies may be faced with dozens of potential reservoir targets.
Seismic and other conventional technologies can say:
"The potential reservoirs are here."
At the decision stage, however, the question changes:
"Which reservoir should I drill?"
It is at precisely this point that FindR is positioned.
FindR is an independent decision-making technology intended to help select the correct reservoir and the correct drilling point from among the targets emerging at the conclusion of the conventional exploration process.
NOT ONLY OIL AND NATURAL GAS
Although FindR technology is described through oil and natural gas, its field of application is not limited to these.
The same approach of the technology can also be applied in the discovery of
geothermal resources, minerals and cold water resources.
FindR's fundamental approach is therefore to assist in identifying the correct target from among numerous potential targets across different natural resources.
Teknolojinin bilimsel ve kurumsal olarak daha güçlü biçimde değerlendirilmesi için uygun sonraki adım, bağımsız kör doğrulamadır. Bu doğrulamada hedef alanlar veya kuyular nihai sonuçlar bilinmeden seçilmeli, değerlendirme kriterleri test başlamadan önce belirlenmeli, veri sızıntısı engellenmeli ve önceden belirlenmiş tüm hedefler sonuç ne olursa olsun değerlendirme setinde tutulmalıdır.
FindR’ın tahminleri daha sonra bağımsız olarak doğrulanmış sondaj ve saha sonuçlarıyla karşılaştırılmalıdır. Böyle bir çalışma; duyarlılık, özgüllük, yanlış pozitif, yanlış negatif, kesinlik, tekrarlanabilirlik ve ekonomik değer gibi performans göstergelerinin bilimsel olarak ölçülmesine imkân sağlayacaktır.
The appropriate next step towards a stronger scientific and institutional evaluation of the technology is independent blind validation. In such validation, the target areas or wells must be selected without the final results being known, the evaluation criteria must be established before the test begins, data leakage must be prevented, and all predetermined targets must be retained in the evaluation set regardless of the outcome.
FindR's predictions must then be compared with independently verified drilling and field results. Such a study will make it possible to measure performance indicators — such as sensitivity, specificity, false positives, false negatives, precision, repeatability and economic value — on a scientific basis.
This workflow clearly demonstrates FindR's role on the oil and natural gas side. FindR is not a technology that replaces the conventional evaluation process; it is a complementary decision support layer operating on the reservoir targets identified by the company.
A. The Company's Conventional Evaluation Process
Using geology, seismic, gravity, magnetic, well data and other conventional methods, the company identifies reservoir areas that may potentially contain hydrocarbons.
B. The Critical Decision Point
"Which of the identified reservoirs contains producible hydrocarbons, and which should take priority for drilling?"
C. The FindR Decision Support Layer
→ Identified reservoir targets → Transmission of the reservoir coordinates to FindR → FindR reservoir evaluation and classification → Target containing no producible hydrocarbons / dry target → Target where a reservoir is present but is predominantly water-bearing or of low economic producibility → Priority reservoir containing producible hydrocarbons → Prioritization of the reservoirs → The company's drilling decision
D. FindR's Core Role in the Process
FindR is not a technology that initiates the conventional hydrocarbon evaluation process, nor one that replaces the geology, geophysics, seismic and other methods used within it. FindR's role in this working model is to evaluate the reservoir targets identified as a result of the work completed by the company and whose coordinates have been transmitted to it, to classify these targets in terms of their content and producible hydrocarbon potential, and to assist in prioritizing them for drilling.
For this reason, in order for FindR to apply this evaluation model, the reservoir targets identified as a result of conventional processes and their associated coordinates must first be transmitted to FindR.
FindR is a data-driven technology aimed at producing a structured evaluation and decision framework in natural resource projects by bringing together past field and drilling results with the measurable characteristics of new targets.
The platform's aim is not to investigate every square kilometre with equal intensity, but to prioritize the areas with the strongest data support through historical analogues, positive and negative field results and multidimensional similarity analysis.
The ultimate purpose of this approach is to direct limited exploration capital towards targets with a scientifically stronger rationale, to support drilling decisions with more data, and to reduce exploration uncertainty.
FindR is a data-driven decision support platform that compares a multi-layered reference data bank — built from oil, natural gas, mineral and geothermal drilling operations whose results are known, together with geological, physical, chemical, surface and soil data — against the standardized multidimensional characteristics obtained from the targets identified for evaluation, that assesses successful and unsuccessful past examples together, and that assists in classifying and prioritizing targets through proprietary data processing and pattern matching algorithms.
