Recently, we have changed the system for points calculation for the Arenas, on Ferion Portal. The new system has been expanded and now the scoring is based on multiple statistics, not just the final population of empires. We believe it's much better, more engaging and fair for players on every level.
What was wrong with the previous system?
To win an Arena in Ferion, an empire must control at least 50% of the total population. Crossing this threshold shows the absolute domination of one player over the others and leaves no doubt as to who the champion is. Once an empire surpasses this limit, it becomes nearly impossible for the remaining players (even through joint effort), to dethrone the victor if he was able to hold more than 50% of population for 24 hours.
This is a good way to determine the winner, but it also means that the remaining players share the other 50% (or, in many cases, much less) of the total population among themselves. This gives the winner a huge advantage in points. They will always have more population (and therefore bigger score) than all other players combined.
HoweverIn reality, the situation is much worse. I have collected statistics from recent games, which are presented in the table below. It shows what percentage of empires ended the game with a population greater than zero. The last column indicates what percentage of players had a significant population share - that is, at least 10% of the winner’s total population.
As you can see, on average only about ~19% of players have any population at the end, and only about 6% manage to maintain it at a significant level. With an average of 48 players, this means that in total 3 empires finish with a satisfactory result in every Arena.
I believe that such a scoring system is unfair to the majority of players. Ferion requires dedicating many hours, but is designed to allow only one person (or in the future, one alliance) to win. What about the other empires, to which players have also devoted a lot of time? Ferion is a complex game where an empire can be developed in many different areas. There is nothing stopping us from rewarding and evaluating player efforts in a much broader range, as long as total points of winning player will keep him on the top of the leaderboard.
What’s behind the new system idea?
The new scoring system is based on three main principles:
The player who wins the Arena (by having the highest population) should be at the top of the leaderboard, which includes additional categories
The system should evaluate the overall performance of players during the game in a broad, but reasonable, scope
The rules for scoring should be simple and clear
The above assumptions (not directly) require an additional condition. Points in each category must be summable to obtain the final score for each player. This is not entirely obvious, let me illustrate with an example:
What happens if we add the number of settled planets to the scoring categories? The winner’s population is counted in tens of thousands, while the number of settled planets is at most a few hundred. Here we have a large difference in scale. Simply adding the number of settled planets to the population will not change the total score. From the other side, adding damage dealt in combat (counted in millions) will have a significant impact on the final scoring.
The solution I applied here was score scaling. It works on a simple principle, separately for each scoring category:
Assign the maximum possible points for that category. For example, 1000.
Find the player who had the highest points in that category in the Arena. That player receives the maximum points (1000 in our example).
Other players receive proportionally fewer points (proportion is not linear, more details later).
Repeat for each category (where a different maximum number of points can be chosen)
In other words, the example of damage dealt always scales to 1000 points, regardless of whether the winning player dealt 10 million or 100 thousand damage. Similarly, the number of settled planets, population, and every other parameter is scaled accordingly.
It is also important that within scaling, we can assign more points to specific categories. This can be compared to a weight system. The mentioned population should have a very high weight compared to other categories to ensure the player with the highest population wins. On the other hand, credits are very easy to obtain, and their significance in the final scoring should not be as high as, for example, acquiring new technologies.
How about details?
Now that I have explained how the new system works, let's dive into the details. What categories are currently scored, and how many points can be earned in each? The image below shows a Radar Chart where you can find the six main categories in which players are scored.
DISCLAIMER: The number of categories and detailed points can change in the future, but the overall rules apply
In most categories, there are several aspects for which players are rewarded. You can find the exact details in the table below:
As you can see, currently it is possible to earn 14,000 points, of which one-third is allocated to population. Such a distribution of weights allows the player who won the arena to maintain the lead. Such a player also performs well in other aspects of the game, so he/she should have good scoring in the rest of the game's aspects. The other categories allow to earn a maximum of 2000 points, except for the economy, which is 500 max.
How is value mapped to score?
Let’s take Population as an example. As I mentioned, the maximum score an empire can earn from population is 5000.
Since the winner had 10,000 population points worth 5,000 scoring points, the simplest approach to convert population for other players is linear scaling. For example, a player with 10% of the winner’s population gets 10% of the points; one with 20% gets 20%, and so on. However, Ferion features exponential growth mechanics, such as technologies ranging in cost from 10 research points to 1 million. Fleet attack strength varies greatly from early small forces to late-game dreadnoughts. This dynamic means a linear scale may inadequately reflect player achievements in all aspects of the game.
These are the reasons why I used an exponential scale to map each value to its corresponding points. It grows quickly at the beginning but slows down in the second half. Below, I present how this works precisely for our example:
The final effect?
As mentioned earlier, under the previous system, only about 6% of players achieved a significant score (at least 10% of the winner's points). How does it look now? On average, 77% of players (compared to 19% before) have a non-zero score, and 38% have earned a significant number of points. This demonstrates a substantial improvement in scoring distribution, making the game more rewarding for a larger portion of players and better reflecting their contributions and effort they put into the game.
