Daphne Growth Rate Ranking
Tracing the Path of Heroes
A detailed ranking of the growth rates and stat individual values (IVs) of all adventurers in Wiz Daphne, calculated using a unique formula.
■Character Ranking Board■
⚙️ Site-Wide Common Restrictions (Upon Database Extraction)
At the point of data retrieval from the database, which serves as the foundation for all features, the following strict limiters are applied.
- 10,000 Entry Limiter: No matter how many entries a character has, a maximum of 10,000 entries per stat will be processed, prioritized by 'Highest Reliability > Newest Update Date'.
- Complete Exclusion of F Rank: Data that has fallen to an 'F' reliability grade due to abnormal value penalties, etc., will be unconditionally excluded from all aggregation and extraction to prevent noise.
📊 Restrictions on the Growth Rate Ranking Screen
In the ranking, restrictions are in place to prioritize maintaining display speed and readability.
- Patterns in the Top 20: Up to 10 individuals
- Patterns from 21st Place onwards: Up to 3 individuals
✦ About Growth Rate Rankings ✦
■ Rank & Growth Rate
This indicates where each stat's growth rate ranks among all candidates.
The amount a stat increases per level up.
Example: If the STR growth rate is 50% (0.5) and the current STR is 10.3, it will become '10.8' when leveling up. (In-game display truncates decimals, so it will still look like 10.)
- On this site, growth rate decimals (e.g., 0.5) may be rounded and expressed as a percentage (e.g., 50.0%).
Rankings are compared between the same characters with the same class.
[+][-]Rank Calculation Algorithm (Details)
Reason for Evaluating by Minimum Growth Rate
The 'Minimum Growth Rate' of each dataset is used as the evaluation axis.
If the maximum value were used, low-accuracy (wide-variance) data would be overestimated, unfairly pushing down the ranks of highly reliable characters. Using the minimum value ensures that as reliability increases through data accumulation, the rank is more likely to rise.
(* Even if your own reliability is high, your rank will constantly update due to fluctuations in surrounding data.)
Exclusion Criteria (Noise Filtering)
To prevent unreliable, uncertain data from unfairly disrupting the ranks of confirmed data, this tool excludes data meeting both of the following conditions simultaneously from rank aggregation as 'noise' (displayed only as 'Equivalent to Rank X').
- Potential for Reversal: Your 'Maximum Growth Rate' exceeds the 'Minimum Growth Rate' of higher-ranked data.
- Subordination of Reliability: Your confidence level (S > A > B) is lower than that of higher-ranked data.
Example (Extraction after sorting by minimum value):
- 1位: S (50.0% ~ 51.0%)
- 2位: S (49.0% ~ 50.0%)
- 3位: A (48.0% ~ 48.5%)
- 4位: S (47.0% ~ 47.1%)
- Excluded: A (46.5% ~ 47.1%)* Maximum value exceeds the top rank, but reliability is lower.
- 5位: S (46.0% ~ 46.5%)
- 6位: S (45.5% ~ 46.1%)
- Excluded: B (45.0% ~ 47.0%)* Same as above.
By filtering out 'unstable data with artificially high theoretical maximums due to low measurement counts' in this way, we achieve a ranking closest to the 'truth' at present, centered around confirmed information like S ranks.
■ About the Rank Calculation Algorithm
To prevent data fluctuations, rankings on this site are based on the 'Minimum Growth Rate'.
Noise Exclusion Logic
To prevent 'uncertain data'—which has few measurements and only an artificially high theoretical maximum—from disrupting highly accurate top-tier data, entries that meet both of the following conditions are excluded from the confirmed rankings.
- Its 'Maximum Growth Rate' exceeds the 'Minimum Growth Rate' of higher-ranked data.
- Its 'Reliability (S~B)' is lower than that of the higher-ranked data.
What does 'Equivalent to Rank X' mean?
This is an 'estimated rank' assigned to data excluded as noise or data with low reliability (C or lower). It does not affect the confirmed rankings, but it allows you to understand where your data's potential stands overall.
Display Example (For Ranks 4 to 7)
- Rank 4 (Confirmed Bracket)
- Equivalent to Rank 4 (Inherits the display rank above, but consumes the 5th place slot)
- Equivalent to Rank 4 (Inherits the display rank above, but consumes the 6th place slot)
- Rank 7 (Confirmed Bracket. Because the 3 slots above are consumed, the next number becomes Rank 7)
■ IV Judgment
When 'XX IV X%' is displayed, it indicates the probability that an individual value bonus (talent) flag is active for that stat.
How Probabilities are Calculated
The current probability is calculated based on the proportion of 'patterns containing an IV flag' among all candidate patterns derived by the system, serving as a rough estimate.
Example: If the following 2 patterns are possible based on the current stats:
Candidate A: 'Allocated BP 8 + STR IV'
Candidate B: 'Allocated BP 9 + No IV'
➔ Since 1 out of 2 applies, it will display as 'STR IV 50%'.
* The actual likelihood depends on factors like the 'growth rate variance' of each candidate, so a displayed 50% does not strictly mean an exact 50/50 chance.
* As more level-up records are added and candidates are narrowed down, it will gradually approach either 100% (Confirmed) or no display (Confirmed No IV).
■ Confidence
The reliability of the predicted results (S to F).
For example, S means an error of 1 BP or less at Level 130, while A means 1 to 3 BP. A lower rank indicates a wider margin of error.
[+][-]Confidence Evaluation Criteria (Details)
Evaluation based on the 'Margin of Error' at Level 130
When the match rate is 100%, the evaluation is based on the 'stat deviation (margin of error) at Lv 130' caused by the variance in the calculated growth rate.
- S : Error of 1 BP or less
- A : Error of 1 to 3 BP
- B : Error of 3 to 5 BP
- C : Error of 5 to 7 BP
- D : Error of 7 to 10 BP
- F : Error of 10 BP or more
* Due to the nature of the calculations, the stats at Level 130 are not finalized until the character actually reaches Level 130.
* As more input data (level-up records) is provided, candidates are narrowed down, and the confidence level will increase towards S.
Match Rate Penalty
The 'Match Rate' with the observed data should inherently be 100%. If it falls below this, it is highly likely there is an error in the data input or prerequisites, resulting in the following penalties:
- Under 100% to 95%: If the error is less than 5, forced demotion to 'D'; if 5 or more, forced demotion to 'F'.
- Under 95%: Forced demotion to 'F' regardless of the error margin.
Penalty for Insufficient Data Input
If a stat has 'less than 5 levels' of input data, and 'there is a tier level bracket (e.g., Unranked Lv 1-20, Lead Lv 21-30) between Level 1 and the highest entered level with zero data points,' the reliability of the prediction significantly drops. Therefore, the confidence of that stat is forcibly downgraded by 2 ranks (e.g., S → B, B → D).
* Even if this penalty reduces the confidence to 'F', as a special exception, it will not cause a 'chain-reaction downgrade' to other stats.
Penalty for Abnormal Total BP
If the total BP calculated from the predicted results of all stats is '11 or more' or '4 or less (*when all stats are entered but the total does not reach the minimum of 5)', it is considered an input error or an abnormal value, and the confidence of all stats is forcibly demoted to 'F'.
* This penalty does not apply to special characters, such as the 'Masked Adventurer', who do not have BP allocation mechanics.
Penalty for Statistical Anomalies (Missing Bondmate Settings, etc.)
The system automatically detects 'impossibly high growth rates' possibly caused by forgetting to configure Bondmates in the tool despite having them equipped in-game. The confidence of that stat is forcibly demoted to 'F'.
Statistical upper limits are automatically calculated based on data collected from all users. Data that significantly exceeds this limit (or data that exceeds the maximum limit for a normal individual despite not having IVs) is subject to this penalty.
* Even if this penalty reduces the confidence to 'F', as a special exception, it will not cause a 'chain-reaction downgrade' to other stats.
* If settings are corrected and the values return to normal, the original evaluation will be automatically restored.
* Data where the IV judgment is uncertain (an intermediate state where the probability is not 0% or 100%) is exempt from the penalty.
Chain-Reaction Downgrade from 'F Rank' in Other Stats
For every stat that receives an 'F' confidence rank, the evaluation of all other stats is downgraded by 1 rank.
Example: If the initial evaluation was 'STR(F)', 'INT(F)', and 'Other(S)', since there are two Fs, the 'S' is downgraded by 2 ranks, making the final evaluation 'STR(F)', 'INT(F)', and 'Other(B)'.