Problems of assessing video game language localization (using the example of S.T.A.L.K.E.R.: Clear Sky)

Burnasheva Liya Rafaelevna Lecturer; Department of English Language; Saint-Petersburg University of the Humanities and Social Sciences Fuchika str., 15, Saint Petersburg, Russia, 192238 burnasheva-liya@mail.ru

Introduction

In modern conditions of globalization, computer games have become an important element of world culture, and their localization is a necessary process that allows adapting game texts for different target audiences. Over the more than eighty-year history of development, both the video game industry in general and language localization in particular have undergone a huge number of changes, however, issues of an objective assessment of its quality still remain unresolved. The relevance of the research topic is due to the fact that despite the objective need of established developers and young companies for high-quality language localization, its level remains quite low, which leads to the need to identify, systematize and formalize criteria for analysis and quality control. In addition, the development of the issue of assessing the quality of translation and localization of video games has a multifaceted significance, affecting economic, linguistic, professional and user aspects. Thus, improving the quality of localization from an economic point of view contributes to the commercial success of video games, strengthening the reputation of developers and publishers on the international market; in linguistics, it allows improving the methodology of studying translation, clarifying criteria for its equivalence and adequacy; from a professional point of view, it promotes the development of the translation industry and improves the training of specialists in this field; for players, it allows deeper immerse yourself in the game world, avoid distorting the meaning and reducing the emotional impact of the game.

To achieve this goal, it is necessary to solve the following tasks::

- correlate the concepts of "translation" and "localization";

- analyze existing approaches to evaluating translation and localization and determine the prospects for their application;

- identify problematic texts of the game S.T.A.L.K.E.R.: Clear Sky and conduct a qualitative and quantitative analysis of the nature and frequency of errors;

- to evaluate the applicability of E. Chesterman's classification of translation transformations for the analysis of video game localization and to identify frequency transformations;

- formulate recommendations for improving the quality of localization of video games based on identified errors and successful solutions.

The research material consisted of 136 original Russian-language texts of the video game S.T.A.L.K.E.R.: Clear Sky (2008) by GSC Game World and their English-language localizations made by the developer company.

The theoretical basis consists of scientific and applied research devoted to the study of translation theory, in particular, the problems of the content of translation activities and its relationship with language localization (Komissarov, 1990; Esselink, 2000; Kazakova, 2002; Sdobnikov, 2018; Sdobnikov, 2018; Achkasov, 2019; Baryshev, Sdobnikov, 2020; Chistova, 2020; Senchuk, 2021; Malenova, 2022), text transformations in translation (Chesterman, 2016), as well as the problem of assessing the quality of translation and localization (Pavlova, 2012; Kashirina and Krykova, 2018; Borisova, 2024). In addition, the study took into account online materials from domestic localization companies, which provided an up-to-date practical perspective on the process of language localization of video games (Krasnovid, 2017; Dubovoy, 2018; Alconost, 2020).

The following research methods are used in this article:

- analysis of scientific and practical sources to identify relevant approaches to defining the basic concepts of research (translation, localization), their main characteristics, as well as to study existing approaches to classifying translation errors and assessing the quality of translation and localization;

- pre-translation analysis to determine the main translation strategies and the direction of further qualitative and quantitative analysis;

- contextual and semantic analysis for the study of linguistic units in the studied texts;

- comparative analysis to compare original and localized texts;

- elements of quantitative analysis to assess the frequency of errors, successful/unsuccessful decisions and the score ratio of texts.

The practical significance of the study lies in the possibility of applying the proposed localization quality assessment system for further research in the field of video game translation and adaptation, as well as in developing methodological recommendations for localization specialists. The study also has the potential to be used in the training of translators and localizers, as the proposed system can be used as a tool for self-control or mutual expertise in the classroom.

Discussion and results

Computer games are multidimensional multimodal works with unique visual and communicative potential, the history of which begins in 1939-1940. During its short existence, this type of multimedia has undergone many changes, developed a complex genre system, and gained a huge audience around the world. Modern computer games are not only a means of leisure, learning, self-improvement and a platform for communication, but also cultural products that are created mainly abroad in English. The objective need to expand the audience and, consequently, increase profits necessitates the need for adaptation to different markets, which is traditionally referred to as localization.

Localization is a multidimensional production process that includes various stages, among which translation work is not in the first place in terms of priority, but perhaps the most difficult and important. However, before the work of translators can begin, there are several more steps to go through, depending on certain conditions. For example, the game can be localized by the developer himself, then you need to choose between machine or crowdsourcing translation, full-time translators, freelancers or a professional localization company. In general terms, the following steps can be outlined on the part of both the customer and the contractor:

So, the customer:

1) analyzes the market and identifies the countries for which localization of the game is required;

2) prepares documents for translation (localization tables, instructions, etc.);

3) if necessary, sign localization agreements and forward the documents to the performers.

Then, when the documents have already been received by the contractor, preparations for localization begin.:

1) in some cases, files to be translated are extracted and organized (if the customer has not done so);

2) a platform (program) is selected for translation, deadlines are set, and work is distributed among translators;

3) Finally, the translation, editing and testing of the finished translation is performed. This step can be referred to as direct language localization.;

4) after that, the actors are cast, the text is dubbed, and the sound is processed;

5) and as a result – the assembly of the localized version of the game, its release and promotion. ^{[8, 14]}.

Thus, linguistic (interlanguage) localization is part of the localization production process. In addition, a distinction can be made by fields of knowledge: localization is the object of research in localization theory (Localization Studies), and interlanguage localization is the object of translation theory (Translation Studies) ^{[13, p. 166]}.

Language localization is designed to transform a foreign language text at the linguistic and cultural levels adequately to the needs and expectations of the target audience and in compliance with the regulatory, political and other features of the host culture. Bert Esselink, the author of one of the first and most popular definitions of localization in the software field, just notes the importance of two aspects – linguistic and cultural: localization is "the process of linguistic and cultural transformation of a product to match the target locale (country / region and language) in which it will be used and sold" ^{[15, c. 3]}. The concept of locale in modern sources is considered not only in the context of region and language: the boundaries of locales do not always coincide with geographical, cultural, linguistic and other boundaries, but rather are determined, among other things, by purchasing power, marketing, political, legal, ethical, gender and other parameters. So, we can talk about different locales: linguistic, legal, cultural, geographical locale, etc. Locale can also be understood in narrow and broad terms: in the narrow sense, it is a set of components that require modification in the process of localization of digital products, and in the broad sense, it is a set of target market parameters that determine the overall strategy of product or brand output. to a new market ^{[1, pp. 82-83]}.

In the software industry, including entertainment, the concept of localization is used synonymously with translation, but a distinction should be made between the two. The simplest and most comprehensive definition of translation belongs to T. A. Kazakova: translation is "the transformation of a message in the source language into a message in the target language" ^{[5, p. 9]}. In other words, it is an interlanguage transformation of the original text using certain translation methods in order to obtain a translated text equivalent to the original one ^{[12, p. 137]}. So, translation is a transformation, the transformation of the original into a product duplicating it in another language (transliterated). Language localization is not just the transformation of one text into another, equivalent to the first, but adaptation, adaptation of the original to various norms of the recipient locale.

V. N. Komissarov also spoke about a phenomenon similar to localization, adaptive transcoding. While translation acts as a full–fledged replacement of the original in another language, adaptive transcoding acts only as a special representation of the original content and represents a combination of two successive transformations - translation and a given adaptation of the text ^{[7, p. 48]}. However, despite the apparent similarity of adaptive transcoding and interlanguage localization, these two concepts cannot be identified ^{[13, p. 168]}; interlanguage localization rather includes translation elements and adaptive transcoding techniques ^{[13, p. 169]}.

Despite the seemingly established practice of using the term "localization" in the context of video game content transformation, debates about the need to distinguish between language localization and translation, and even more so the coexistence of the two terms within the framework of translation theory, do not subside. Since the process of language localization involves, in fact, the translation of the text of the game from one language to another, opinions are expressed about the redundancy of the term "localization". Thus, V. V. Sdobnikov is perhaps the main opponent of this term and considers naive the opinion of some theorists and practitioners that translation is only the reproduction of cognitive information in the same form as in the source text, and localization is a separate process aimed at cultural adaptation of the translated text, depending on the specifics of the translated text. the text or the importance of the cultural factor within its framework. Meanwhile, in the actual translation, cultural adjustment is inevitable and indisputable for the recipient – the difference is only in the degree of severity of this adjustment ^{[17, p. 1495-1496]}, because translation is both a linguistic and cultural transfer of texts in the original language into the linguistic and cultural space of the target language ^{[2, p. 19]}. Thus, we can say that high-quality translation also includes pragmatic adaptation and is fundamentally no different from localization ^{[11, p. 76]}.

However, in the context of at least the video game industry, it seems quite appropriate to make such a distinction, since we are talking about different equivalences: direct translation implies the equivalence of the message content, and localization implies the equivalence of a unique user experience within the native linguistic culture ^{[9, p. 73]}. It is also important to point out the sometimes occurring difference in the completeness of the text that ends up with the user. As I. V. Borisova notes, translated content is not always fully integrated into the final product ^{[3, p. 193]}. Such changes are not uncommon for different types of translation itself (abstracting, annotating, literary retelling, etc.), however, in the case of language localization, another motivation applies: the content of the texts themselves or parts of the original non-textual content may be unacceptable to the end user, and therefore they are replaced or deleted in the final version. In other words, language localization can affect not only the text component, but also the game content itself, which can be partially changed, replaced or deleted. When translating directly, the effect of the translation product on the content and form of the source text is not assumed, and the differences from the original in volume and completeness are due to the purpose of the translation.

The controversial position of interlanguage localization leads to attempts to identify its fundamental differences from other types of translation activities. For example, E. V. Chistova provides a comparison of literary translation, industry translation and interlanguage localization according to the following parameters:

- consideration of cultural peculiarities;

- consideration of legal aspects;

- consideration of market needs and marketing strategies;

- consideration of stylistic features;

- consideration of visual elements of polycode texts;

- accounting for auditory and media components of multimodal texts;

- adaptation of jokes;

- programming language recognition.

Thus, interlanguage localization satisfies all of these parameters, while artistic and industry translation satisfy only some of them (artistic translation takes into account stylistic features and adapts jokes, and also rarely takes into account cultural features and visual elements of polycode texts; industry translation takes into account only stylistic features and visual elements) ^{[13, p. 170]}.

Based on the above, the following main differences can be formulated between localization as a production process, language localization and translation proper:

1. Localization as a production process:

- it consists of many stages, only one of which is language localization.

2. Language localization:

- includes the actual translation, but goes beyond it by changing the texts in accordance with the legal, marketing and other requirements of the target market;

- it is involved in working with multimodal texts, and therefore, in addition to linguistic and cultural features, it takes into account audiovisual components.;

- strives for the equivalence of a unique user experience;

- it can lead to changes or shortening of the translated text and even the game content itself for reasons of impropriety for the target audience. In some cases, it may affect the game creation process.

3. Translation:

- focused on conveying the content of the source text without changing its shape and structure;

- strives for the equivalence of the message content;

- does not lead to changes in the original text and non-text components;

- the differences in volume and completeness from the original text are due to the purpose of the translation.

Thus, localization covers the entire process of adapting a product to the market, language localization is precisely working with texts as part of this process, and translation is the transformation of text without additional changes in the content structure.

Whatever the attitude to language localization in the academic and professional environment, there is an objective need to evaluate its quality and a certain list of recommendations and requirements for the assessment system. But first of all, it should be noted the following complicating feature of the localization of a video game product is the frequent inability to "see the full picture", in other words, forced work without context details. Therefore, the assessment of the quality of language localization can take place before the release of the game (pre-release assessment), and after the release of the game, taking into account the actual feedback from the players (post-release assessment).

A single generally accepted approach to assessment has not yet been developed: experts propose different parameters and their characteristics ^{[6, p. 124]}. However, there are several recommendations for developing such a system. First, it is necessary to understand from which point of view the analysis should be carried out: linguistic (in terms of translation studies) or social (from the perspective of the players). Secondly, an effective translation quality assessment system should be based on a clear classification of translation errors, which will identify deviations from generally accepted translation standards. Nevertheless, due to the high variability of information transmission methods and the creative approach to translation, it is impossible to name the correct or erroneous translation option. In addition, it is impossible to unambiguously determine the nature and time of the error ^{[6, p. 125]}. Among the requirements for the translation quality assessment system, experts identify the following: impartiality of assessment by a disinterested person, universality, an understandable and accurate classification of translation errors, specific evaluation criteria, a clear scale of assessments, and applicability in practice ^{[6, p. 126]}. Additionally, you can mention the technical requirements for observing the number of characters (for user interface elements, subtitles; they may vary depending on the limitations of the engine and regional standards) and lipsink.

However, as for impartiality, it seems that any assessment cannot be completely objective: even resorting to formalized assessment methods cannot completely eliminate the subjectivity of the evaluator, and also does not solve the problem of subjectivity of perception on the part of the player. For example, the evaluator may have nothing to do with the location being evaluated and the project as a whole, however, the assessment will be influenced by personal linguistic and stylistic preferences, gaming experience, and cultural differences. In addition, a pre-release localization assessment cannot be performed by a person who is not familiar with the general localization process, as this assessment will be unprofessional and largely superficial. And if the localization was evaluated by a specialist and published, it, despite being correct, may be considered erroneous due to inconsistency with the expectations of the audience. Thus, an absolutely unbiased assessment of the quality of video game localization seems impossible, but this can be approached by involving a qualified independent expert, using formalized evaluation criteria, taking into account existing localized parts of the game and other materials on which the game can be based, as well as analyzing the characteristics and expectations of the target audience.

Speaking about the errors of language localization, experts from Inlingo Games name the following key errors that determine its quality:

1) actual translation errors (untranslated text, erroneous translation, unauthorized additions by the translator);

2) language errors (at the level of grammar, spelling, punctuation, style);

3) disregard of recommendations (non-compliance with instructions, rules of work on the project, comments of the editor, project manager or client);

4) errors in terminology (interface, glossary of the project, heterogeneity);

5) technical errors (non-compliance with limits, etc.);

6) nomadic semantic errors (due to a "bridge" translation, i.e. performed using a third intermediary language) ^[4].

It is also appropriate to cite the typology of translation errors proposed by V. N. Komissarov.:

1) gross distortion of the original content, misinforming the recipient, which occurs as a result of an erroneous understanding of the text in English;

2) inaccurate rendering of the meaning of the original, without distorting it completely;

3) a decrease in the quality of the translation text as a result of deviations from stylistic norms, the use of uncommon vocabulary, etc., leading to a distortion of the meaning of the original;

4) violation of mandatory norms of translation, which do not affect the equivalence of translation, but call into question the general linguistic culture and literacy of the translator ^{[7, pp. 241-245]}.

A.V. Pavlova has developed a more detailed system for assessing the quality of translation – a multidimensional matrix of criteria for evaluating translation, consisting of 13 requirements for equivalence due to the original discourse; 6 requirements for equivalence due to the host discourse; parameter of equivalent / unequal functionality, impoverishment / enrichment; parameter of the degree of freedom of the translator. This detailed matrix describes the aspects that affect the completeness and, ultimately, the quality of the translation, for which the translation can be analyzed by comparing and contrasting it with the original. This matrix is designed, if not to rid the assessment of subjectivity, then at least to minimize it and increase the fairness of the assessment ^{[10, pp. 36-38]}.

Such typologies of errors can be supplemented with point designations in accordance with the "severity" of each of the points and based on them, a quantitative (digital) analysis of the quality of the translation can be carried out (as described, for example, by V. N. Komissarov ^{[7, p. 242]}). The distribution of points may vary depending on the complexity of the project and the personal experience of the evaluator, which is again associated with subjectivity, however, this approach allows you to systematize the assessment and provide a greater level of objectivity than, for example, in the absence of scoring designations.

The error typology proposed by Inlingo Games is suitable for both pre-release and post-release localization quality assessment, but it is more likely focused on the localization translators themselves due to the criterion of recommendations and instructions for working with the project. A.V. Pavlova's matrix is excellent for deep post-release linguistic analysis of translation, however, due to the large number of elements and obviously high labor and time costs, it cannot be an effective tool for pre-release evaluation of the quality of video game localization, a process that is often extremely time-limited. Therefore, at the stage of pre-release testing and localization quality assessment, it seems sufficient to conduct a general assessment of the quality of the translation, for which you can use the error system proposed by V. N. Komissarov; for post-release evaluation, it is also possible to base on the typology of errors by V. N. Komissarov.

In the further analysis, the focus is on working with text content directly, without taking into account technical and other organizational aspects due to the objective impossibility of immersing oneself in the real conditions in which work with the text of a video game took place. The result is the following translation quality assessment system, in which points are assigned to each category, and the maximum amount of points is 7:

1. Gross errors that lead to a serious distortion of the meaning of the original (3 points). The most critical and multidimensional type of error, therefore, it is assigned the highest number of points in this system. This type of error consists of six aspects:

- violation of denotative equivalence;

- violation of pragmatic equivalence;

- illegal addition or omission of information;

- grammatical error;

- lexical error;

- incorrect reading or understanding of the original.

(An additional 3 points can be equally divided between these aspects for a more quantifiable analysis; however, this study does not provide for such a division.)

2. Minor errors, inaccuracies in conveying the meaning of the original, not related to serious distortions (2 points).

3. Stylistic and other shortcomings that lead to a decrease in the quality of translation (1 point).

4. Violation of mandatory rules and regulations (1 point).

Within the framework of this system, localized texts with a maximum score of 7 can be considered an unequal and inadequate translation, and texts with a score of 0 can be considered an adequate translation that takes into account the original pragmatics and norms of translation. The above system can be used both to evaluate the translation as a whole and private components: the translation of character names, toponyms, interface elements, etc.

According to the described evaluation system, the official English-language localization of the texts of the video game S.T.A.L.K.E.R.: Clear Sky, made by the developer studio GSC Game World, was analyzed. For quantitative analysis, the classification of translation transformations by E. Chesterman was also applied, which identifies the following groups and transformations: syntactic (literal translation, tracing paper, transposition, etc.), semantic (conversions, synonymy, antonymy, hypo-, hyperonymy, etc.), pragmatic (cultural filtering, changing the level of explicitness, partial translation, etc.) ^{[16, pp. 91-109]}.

As a result of the analysis of the texts of the game S.T.A.L.K.E.R.: Clear Sky, the following was revealed: of the 136 texts of the game included in the sample, one third (46 texts) are an adequate translation (received 0 points), two thirds (90 texts) are recognized as not quite adequate and equivalent to the original.

The category of gross errors accounts for 46% of all erroneous translations, of which 57% (25 texts) are illegal additions or omissions of information, about 15% each are violations of denotative equivalence and lexical errors (7 and 6 texts, respectively); the remaining 13% are violations of pragmatic equivalence (3 texts), incorrect reading original text (2 texts) and grammatical errors (1 text).

The ratio of the other error categories is as follows: non–gross errors – 31% (29 texts), reduced translation quality - 21% (20 texts), violation of mandatory standards – 2% (2 texts).

It is worth illustrating the above calculations. Of the third texts that received 0 points in the described assessment system, the following examples can be given:

1. Semantic transformation:

«...Well, we welcome everyone: some with bread and salt, some with just salt, and some with lead" / "...We welcome them all, only some get the open-arms treatment, and some the feet-first kind." In the localization of the replica, the food component of the Russian phraseology ("bread and salt") is replaced and metonymy ("to meet with lead") into an idiom with somatic components in English ("open-arms treatment", "feet-first treatment"), which retains the necessary connotation and gives dynamics in translation, as it uses other motivation mechanisms.

2. Syntactic transformation:

"Welcome to Yantar, or Dzintars! That's what we call him among ourselves. A joke of humor, yes" / "Welcome to Yantar lake, where you can catch a hake. Ha-ha! You get it? It rhymes! Ha-ha!" In localization, a reference to the Latvian manufacturer of perfumes and cosmetics Dzintars (translated as Amber), which was irrelevant to the recipient, was compensated by the introduction of a rhyming pun, reinforced by a semantic discrepancy: the hake marine fish is not found in fresh lake water.

3. Pragmatic transformation:

"... There are a couple more abandoned villages, you can choose a house for yourself. With or without neighbors, it's up to luck." / "Oh, there are also a couple of deserted villages where you can settle down in a home of your choice. If you’re lucky you’ll even get an invisible roommate with tentacles and all". Information is added to the localization, which enriches the remark with irony and unobtrusively warns the player about upcoming dangers.

It is necessary to illustrate the errors in the above classification. So, the following examples of the most frequent gross errors (46%) can be identified:

1. Illegal addition of information (57%):

"Have you heard what the "dolzhans" are bragging about doing to us? Pulverize is still a minimum. To be honest, this case bothers me..." / "Have you heard what the Duty boys are saying they'll do to us? They’re promising to grind us to dust, and that’s just for starters. I’m pretty worried about the whole thing... Especially given that they’re said to have re-equipped themselves with some top-notch hardware". An entire sentence has been added to the localization, which was not stated in the original and is not logically deducible. Adding it does not carry any important information for the player, moreover, it can be disorienting and, moreover, unnecessarily increases the volume of the text.

2. Violation of denotative equivalence (16%):

"The Swamps used to be a peaceful place: no one wandered here, there was nothing interesting for swag seekers here" / "The Swamps used to be a peaceful place. No one disturbed us, no one came looking for artifacts". There is a violation of reference correlation in localization, which can lead to disorientation of the player: "swag", including artifacts, weapons and other objects, has turned into "artifacts" in translation, which actualizes only one of the components of the original referent. However, further on in the texts of the game, one can also see the lexemes "loot" and "looting", so it is not entirely clear why the translators, having the necessary means, did not bring the texts to lexical uniformity and denotative accuracy.

3. Lexical error (14%):

"What else... Well, I heard they found caches of old machinery, equipment and other stuff in the Landfill, which was buried back at the time of the accident in 'eighty-six." / "Let's see what else... They found a cache of old vehicles, equipment and other goodies at the Garbage – stuff that dates back all the way back to 1986, the time of the original accident". In localization, there is a distortion of the text with extralinguistic cultural and historical realities: it is better to convey the real cemetery of infected equipment with the terms "cemetery of radioactive vehicles" or "the Chernobyl car graveyard", since the "cache" used, although it really has the meaning of "cache", rather means a certain hiding place, which does not fully correspond to the original the meaning.

Rough mistakes can be illustrated with the following text (31%):

"Your first time in the swamps? Then listen to me. First of all, you need to be afraid of other people, and only then monsters and swampy places. Stick to the reeds, then you'll live longer" / "Your first time at the Swamps? Then you listen up. Above all else, it’s the humans you gotta watch out for. Monsters and quicksand are a distant second. Stick to reed beds and you’ll live longer". In localization, the combination of "marshy places" in this case (as well as the lexeme "marshes" in others) is rendered as "quicksand", which, due to the semantic component of "sand", can lead to a false association of marshes – flat, grassy marshes – with quicksand. It would be appropriate to replace "quicksand" with the synonyms "bog" or "marshes" (the latter is even used in other texts of the localization in question). It can be seen that errors of this type, although they do not lead to serious distortions, still entail an inaccurate perception of the original meanings.

The following striking examples (21%) can be distinguished from stylistic and other shortcomings:

The name of the Gulp of Freedom bar is literally translated as "Gulp of Freedom". The denotative equivalence is preserved, but in the context, a pun is implemented based on the name of the bar: "... And if you go to the Gulp of Freedom, you'll swallow a lot, I promise you that. They'll warm you up and feed you there, or they'll give you a job." In translation, this pun is lost, and the phrase reads as follows: "... but if you want to take a well-earned break, visit the Gulp of Freedom. The Gulp is known throughout the Zone for its generous servings, and it’s a good place to pick up a job if you’re up for it". The translation can be made brighter and enriched with a pun by introducing the multi–valued lexeme "shot": this is both a shot and a sip (that is, a "shot glass" is a "shot glass"). The resulting combination of "Shot of Freedom" can be perceived in two ways: "A sip of freedom" and "A Shot of Freedom", while both meanings are relevant in the context of the game.

The ambiguous "Zone" in Russian (the Exclusion Zone and the slang name of the penitentiary institution) does not match semantics with the English-language "Zone", which is why the pun in translation is lost and even the meaning of some phrases is lost. For example, to the question: "What can you tell us about the Zone?" – the character responds as follows: "Which zone should I tell you about? Oh, about this one... Well, this is not a zone at all, but an earthly paradise for a real thief." Translated, this remark takes the following form: "Which Zone are you asking about? This one? This is just a children’s playground compared to some other joints I’ve been in. A real paradise for a thief". Additionally, due to the lexeme "joint", the Exclusion Zone acquires a negative connotation. Objectively, it is impossible to keep both referents in translation, so it makes sense to replace one of them, while preserving the pun and the character's speech portrait: "Which Zone are you asking about? A Friend Zone? Ha-ha! Well, listen, this Zone that we’re in is a real paradise – a happy skimming ground for a thief. Ha-ha!"

As an example of a violation of norms and usage, the localization of the following remark (2%) can be cited:

"Oh, I'm tired of dragging swag out of the Zone, fighting with creatures and bandits for a place in the sun. I took a little break here." / "Man, I'm tired. I’m tired of trying to get loot out of the Zone, tired of fighting monsters and tired of competing with bandits for a place under the Sun. I’m taking a little break here". The calculated "a place under the Sun" should be replaced with the grammatically correct "a place in the sun".

The given examples from a continuous sample represent a "slice" of localization, that is, it is possible to more or less confidently extrapolate the calculations to the entire localization of a part.

The score ratio among the 136 texts of the localized version of S.T.A.L.K.E.R.: Clear Sky according to the given rating system is as follows:

- 33% (46 texts) received 0 points, which corresponds to an adequate translation without errors.;

- 33% (45 texts) received 3 points and 19% (26 texts) – 2 points, which roughly corresponds to the ratio of gross errors and inaccuracies.;

- 13% (17 texts) were rated at 1 point, which corresponds to stylistic and other shortcomings and violations of mandatory norms and usage of the translating language;

- only about 2% of the texts (2 units) received a total of 4 points, which corresponds to a combination of gross errors and stylistic inaccuracies.

Taking into account the obtained score ratio, the English-language localization of S.T.A.L.K.E.R.: Clear Sky can be assessed as average in quality due to the presence of a significant number of gross distortions and inaccuracies, as well as the loss of original expressiveness. It is important to note that this assessment echoes the criticism of the English-language localization from the players, who also point to stylistic inconsistencies and a weak rendering of the atmosphere of the original.

The ratio of translation techniques and the success of the result in localization as such has not been revealed, however, it can be unequivocally said that adding or omitting information (30% of errors versus 7% of adequate translation) and using synonymy (29% of errors versus 20% of adequate translation) lead to an overwhelming number of errors and inaccuracies in translation. Among the other techniques, there was no significant discrepancy in the number of adequate and inadequate (erroneous) translated texts. In this regard, the conclusion is as follows: there is no direct connection between the translation techniques used and the success of the translation result. However, it is obvious that arbitrary interference with the content of texts themselves, as well as manipulation of the semantic component at the word level, can and will lead to violations of the original meaning, which in turn, due to the interactive and multimodal nature of the video game, entails a number of other complications for the player.

Summarizing the quantitative analysis, it is worth noting the potential causes of erroneous translation. Thus, gross errors and inaccuracies resulted from a rather free-handed approach to translation: many texts underwent unjustified additions or omissions of information, as well as the replacement of original units with synonyms that were inappropriate in the context. Earlier it was stated that the localization into English was handled by the developer company; presumably, this was the reason for such manipulations with the content of the texts: the localizers were immersed in the details of the plot, were in context, and showed excessive creativity in some places (and in some places, on the contrary, they did not work out the original enough, probably also because of confidence in the deep knowledge of the material). That is, being in the context does not necessarily lead exclusively to positive results; sometimes, on the contrary, it entails harmful inconsistencies with the original on many levels, from lexical and stylistic to cultural. Violations of norms and practices are less common, since they are easy to identify and eliminate, nevertheless, their presence, albeit in small numbers, indicates, for example, an unwillingness to consult reference materials due to the same confidence in their knowledge.

In general, the English localization copes with the main tasks, but it looks inauthentic, dry and even inanimate. Although phraseological phrases, colloquial formulas, and other marked expressions are conveyed by formal logical equivalents, which makes it possible to understand the original meaning, dynamic equivalence is often absent. This is often accompanied by an additional increase in the lexical composition in the translation, which makes the text heavier, deprives it of elegance and blurs the content. The result is a collection of stylistically rather disparate "semi-academic" texts that convey the speech of well-educated individuals (which actually runs counter to the general context of the game) interspersed with conversational elements. We can explain this outcome: localizers are not native English speakers. Nevertheless, nothing seems to have prevented linguistic testing involving native speakers or hiring a consultant to help with localization, since one should not forget that in video game localization, accurate transmission of factual information is often inferior in importance to achieving the necessary pragmatic effect ^{[6, p. 125]}.

Conclusion

Thus, the conducted research allows us to come to the following conclusions:

1. The correlation of the concepts of "translation" and "language localization" showed that the latter covers a wider range of tasks, including adapting content to the cultural, technical, legal and other features of the target audience.

2. In the course of the study, existing approaches to assessing the quality of translation and localization were considered, and an adaptation of V. N. Komissarov's error system, supplemented with scoring marks, was proposed to increase objectivity in assessing the quality of localized texts.

3. Qualitative and quantitative analysis of the texts of the game S.T.A.L.K.E.R.: Clear Sky has demonstrated that the most frequent errors are the illegal addition of information, violation of denotative equivalence, stylistic inconsistencies.

4. E. Chesterman's classification of translation transformations has the potential to be used for such an analysis. It allows us to identify some patterns (for example, liberties in working with the original content, leading to inaccuracies and errors in translation), but it is impossible to draw an unambiguous conclusion about the relationship between the transformations used and the success of their outcome.

5. Based on the identified errors and successful solutions, the following recommendations for assessing the quality of language localization are formulated, which determine the prospects for further study:

• studying the differences in approaches to localization in different markets and identifying universal and specific problems;
• development of a unified methodological approach to assessing the quality of localization of video games, separate approaches for different genres or markets;
• formulation of a clearer concept of translation error and its varieties;
• the introduction of the following criteria into the evaluation system for a deeper analysis: "compliance with existing realities", "compliance with the expectations of the target audience", "naturalness of speech" / "authenticity of the text", "ease of perception", "uniformity of terminology", "consideration of the number of characters", "convenient length of the displayed subtitle", "compliance with lipsink" and so on .;
• compilation of tables/documents with classification and description of the most frequent errors in each genre for consideration in future projects;
• It is possible to implement an automated assessment using artificial intelligence.;
• involvement of native speakers in conducting pre-release testing of localized versions of games.

Thus, the study confirms the need to develop standardized methods for assessing the quality of language localization, however, following K. Y. Kashirina and I. V. Krykova, at the moment only one conclusion can be reached: it is difficult to imagine a system for assessing the quality of translation and language localization that would be completely clear, objective and universal ^{[6, p. 129]}. Even when it appears and takes into account all the various factors, the decisive word will remain with the players.