Aydın Tiryaki and DeepSeek
Introduction
Asking an AI assistant a question about which teams will be relegated from a sports league after the final week’s matches—a league for which it has no real-time, structured data. This seemingly simple request has, in fact, turned into a perfect case study that tests the capabilities, limits, and ethical positioning of modern large language models (LLMs) and the ecosystem of tools that surround them. This article examines this dialogue under a microscope, presenting an analysis of the process in terms of AI methods and drawing inferences about the future of such interactions.
1. Anatomy of the Task in Terms of AI Methods
The user’s question represents not a single operation for an AI assistant, but several nested cognitive tasks. Successfully managing these tasks requires the orchestration of different AI components that make up the system.
1.1. Information Retrieval and Retrieval-Augmented Generation (RAG)
The assistant’s first step in this task is accessing current information. Large language models are limited to the knowledge contained in their training dataset and do not incorporate a “world model.” Therefore, for real-time data like the 2026 Süper Lig standings, the model must necessarily consult external sources.
This is where the Retrieval-Augmented Generation (RAG) architecture comes into play. The system converts the user’s question into a search query and scans current online resources (sports news, league table websites, Wikipedia). However, the critical issue here is the unstructured nature of the information. The league table is a table, while the tie-breaker calculations are a function of past match results. When collecting this scattered information, the assistant, much like a human, seeks a “consensus of accuracy” by cross-referencing information from different sources.
1.2. Layers of Reasoning: From Deterministic to Probabilistic
The raw data obtained triggers reasoning processes at different levels of complexity:
- Deterministic Logical Inference: The assistant’s ability to determine by looking at the league table that “Fatih Karagümrük and Kayserispor are mathematically relegated” is a rule-based inference. It can reach this conclusion on its own through a simple logical operation (if-else) using the variables of remaining matches and point differences.
- Constrained Combinatorial Reasoning: The information that “Antalyaspor will definitely be relegated if they fail to win” is an analysis of a slightly more complex but still limited probability space. The model can confirm this conclusion by calculating the maximum points the opponents could achieve.
- Complex Planning Under Incomplete Data (What Could Not Be Done): The real challenge is calculating the three-way and four-way tie-breaker scenarios. This is not just a reasoning exercise but also a data collection and processing (tool use) problem. The system would need to find the scores of all relevant matches played throughout the season, interpret the ranking criteria in the TFF regulations (such as head-to-head record first, then overall goal difference) as an algorithm, and use a tool (such as a Python interpreter) to execute this algorithm. The system’s failure here reveals one of the greatest limitations of current AI assistants on their path to becoming “end-to-end autonomous researchers”: the lack of a fully mature planning and tool-use capability to dynamically integrate and process fragmented data sources.
1.3. Source Reliability and the Modeling of Uncertainty
The system’s turn to ready-made analyses in sports media, instead of its own incomplete calculation, demonstrates AI’s strategy for dealing with uncertainty. This is not an “error” but a rational decision. The assistant delegates a task it cannot perform itself to an expert subsystem (in this case, a published analysis by a human expert). The ethical and practical responsibility here is to communicate this delegation transparently to the user. By saying “according to analyses, the highest probability is…”, the model must emphasize that this information is not a deterministic truth it produced itself, but a probabilistic estimate taken from a competent source. This is the most critical example of an AI’s epistemic humility and transparency towards the user.
2. Strengths and Weaknesses of the Approach
An evaluation of the case in terms of AI methods reveals the following table:
Strengths:
- Source Synthesis: The ability to transform scattered information from multiple sources into a coherent narrative.
- Uncertainty Management: The capacity to stop at the point of inadequacy, redirect the task to a more competent source, and express this situation transparently.
- Inference Hierarchy: The ability to make simple logical inferences and distinguish complex ones.
Weaknesses and Limitations:
- Lack of End-to-End Tool Use: The inability to independently construct the necessary tool chain (querying a database + running an algorithm) for a complex calculation.
- Dependence on Unstructured Data: The lack of direct API access to a specific dataset like head-to-head records makes the process difficult.
- The Risk of “Blind Trust in the Source”: The verification mechanism ultimately relies on the mutual confirmation of sources. If the entire sports media makes a common calculation error, the model has no chance of detecting it.
3. Recommendations and Cautions for Users: A Guide to Working with AI
This case also holds important lessons for users of AI assistants:
- Deconstruct the Anatomy of Your Query: The question “Who will be relegated based on today’s match results?” is actually a chain of “data collection + rule application + scenario simulation.” Predicting where in this chain the AI might struggle allows you to ask better questions (like “Can you first list all the matches these teams played against each other?”).
- Question the Degree of Certainty: When the assistant uses phrases like “definitely,” “with the highest probability,” or “according to analyses,” this is not just a language preference but a signal about the nature of the information provided. It is healthiest to catch these signals and, especially when “according to analyses” is mentioned, to question the primary source of those analyses.
- Use the Assistant as an Auditor: Your question, “Wouldn’t it have been better to compile all the match results and do it yourself?” is an excellent audit question. It forces the assistant into an internal accounting of its capabilities and process. Asking these kinds of “process audit” questions at the end of every complex task increases the quality of your relationship with the AI.
- Do Not Trust Blindly; Build a Verification Chain: No AI assistant is the ultimate and flawless authority. Especially in events with rapidly changing and complex rules (sports competitions, election results, financial markets), see the AI as an intelligent assistant and make the information it provides the first link in your own final verification chain.
Conclusion
This dialogue, which began with a question about relegation in the Turkish Süper Lig, became a live simulation of how an AI assistant steps out of the laboratory and grapples with a messy, real-time, and complex real-world problem. This interaction has shown that the power of AI lies not only in finding patterns in massive datasets but also in recognizing its own limits and expressing these limits transparently to the user.
The greatest lesson to be drawn from this is that the future of human-machine collaboration lies not in the machine’s ability to imitate humans, but in a model of collective intelligence where human oversight and reasoning capabilities complement the machine’s speed of information access and synthesis. The real issue is not whether an AI can perform a calculation on its own; it is whether the user possesses the conscious skepticism to question the nature, limits, and risks of the information the assistant provides. This case proves once again that this conscious skepticism will be the most fundamental literacy skill of the AI age.
APPENDIX
The Final Knot in the Süper Lig: All Scenarios on the Relegation Battlefield
Aydın Tiryaki & DeepSeek
The 2025-2026 Süper Lig season is witnessing a breathtaking relegation race with just one week remaining. With Galatasaray having already claimed the championship and Fenerbahçe securing the runner-up spot, all eyes have turned to the lower reaches of the table . While Fatih Karagümrük and Kayserispor have already been cut adrift from the league , four teams are locked in a fierce struggle for survival, with only the third and final relegation spot remaining. In this article, we take a closer look at the standings heading into the final week, the teams’ fixtures, and all possible scenarios that arise under the TFF tie-breaker rules.
1. Current Standings Heading into the Final Week
According to the official TFF league table and sports media sources, the situation of the teams in the relegation zone as of May 16, 2026, is as follows :
| Pos | Team | P | W | D | L | GF | GA | GD | Pts |
|—–|——|—|—|—|—|—|—|—|—|—|
| 13 | Eyüpspor | 33 | 8 | 8 | 17 | 30 | 45 | -15 | 32 |
| 14 | Kasımpaşa | 33 | 7 | 11 | 15 | 32 | 49 | -17 | 32 |
| 15 | Gençlerbirliği | 33 | 8 | 7 | 18 | 33 | 47 | -14 | 31 |
| 16 | Fatih Karagümrük | 34 | 8 | 6 | 20 | 31 | 54 | -23 | 30 |
| 17 | Antalyaspor | 33 | 7 | 8 | 18 | 32 | 55 | -23 | 29 |
| 18 | Kayserispor | 33 | 5 | 12 | 16 | 25 | 61 | -36 | 27 |
Note: As the table shows, the relegation of Fatih Karagümrük and Kayserispor has already been confirmed mathematically .
2. Final Week Fixtures
According to the program announced by the Turkish Football Federation, the 34th and final week’s matches are as follows:
Sunday, May 17:
- Fenerbahçe vs. Eyüpspor (20:00)
- Kasımpaşa vs. Galatasaray (20:00)
- Antalyaspor vs. Kocaelispor (20:00)
- Trabzonspor vs. Gençlerbirliği (20:00)
The fact that all matches will kick off at the same time eliminates the possibility of teams playing with knowledge of other results and pushes the tension to its peak.
3. How Does the Tie-Breaker System Work?
Before moving on to the scenarios, let us recall the ranking criteria applied by the TFF in the event of a points tie :
If two teams finish level on points:
- Head-to-head record (points superiority in matches between them)
- Head-to-head goal difference
- Overall goal difference
If three or four teams finish level on points:
A hypothetical mini-league table is formed from the matches the teams in question played against each other during the season. The ranking in this group determines the final standings in the league. The team finishing last in this group is relegated.
4. All Scenarios
Scenario 1: Antalyaspor Fails to Win
If Antalyaspor draws or loses against Kocaelispor, its points will remain at 30 or 29. In this case, regardless of other results, Antalyaspor is relegated.
This is the clearest part of the equation. Since Eyüpspor and Kasımpaşa already have 32 points each, if Antalyaspor cannot reach this threshold, no further calculations are necessary.
Scenario 2: Antalyaspor Wins
Let us assume Antalyaspor defeats Kocaelispor at home. In this case, its points will rise to 32. Now the calculations change entirely.
2.1. Four-Way Tie-Breaker Scenario
In the event that Antalyaspor, Eyüpspor, Kasımpaşa, and Gençlerbirliği all finish the season on 32 points, the four-way head-to-head table would see Kasımpaşa relegated.
2.2. Three-Way Tie-Breaker Scenarios
In a three-way tie scenario involving Antalyaspor, Gençlerbirliği, and Eyüpspor, Eyüpspor is calculated to be the relegated team.
Similarly, in another three-way tie scenario where Kasımpaşa, Antalyaspor, and Eyüpspor finish level on points, Eyüpspor would again be relegated.
In the event of a points tie between Antalyaspor, Gençlerbirliği, and Kasımpaşa, Kasımpaşa is identified as the team to be relegated.
2.3. Two-Way Tie-Breaker Scenarios
It is noteworthy that Antalyaspor holds the head-to-head advantage over all its rivals. Therefore, if it enters a two-way tie-breaker against any team, it is in a favorable position.
5. Final Situation of the Teams
Antalyaspor: The team in the most critical situation. They must defeat Kocaelispor at home. A draw or loss would mean they are mathematically relegated.
Eyüpspor and Kasımpaşa: Both are on 32 points. Even if they win their own matches, they may still find themselves subject to tie-breaker calculations depending on other results. The fact that Eyüpspor faces Fenerbahçe and Kasımpaşa faces Galatasaray in particular makes their task significantly more difficult.
Gençlerbirliği: On 31 points. They will play away against Trabzonspor. A victory would take them to 34 points and guarantee their safety. In the event of a draw or defeat, however, tie-breaker calculations could come into play depending on the results of the other matches.
Conclusion
The final week of the 2025-2026 Süper Lig season is locked in on the match Antalyaspor will play at home against Kocaelispor. A victory for the Mediterranean side would take the matter into multiple tie-breaker scenarios and result in the relegation of either Kasımpaşa or Eyüpspor. A loss of points for Antalyaspor, on the other hand, would directly result in their own relegation.
The picture that emerges from all the scenarios is this: of the four teams on the relegation line, only one will bid farewell to the league. The identity of that team will become clear with the final whistle of the four critical matches kicking off at 20:00 on Sunday, May 17.
| aydintiryaki.org | YouTube | Aydın Tiryaki’nin Yazıları ve Videoları │Articles and Videos by Aydın Tiryaki | Bilgi Merkezi│Knowledge Hub | ░ Virgülüne Dokunmadan │ Verbatim ░ | ░ Yapay Zekanın Türkiye Süper Ligi’nde Küme Düşme ile İmtihanı │AI on Trial: Relegation in Turkish Super League ░ 17.05.2026
Aydın Tiryaki 