Reports on online games and their confirmations (Spring 2009 season).

A total of 472 games were played in the Season, with 111 active players (compared to 470 games and 86 players in the winter season). On average, each player participated in 8-9 games, slightly less than in the winter season.

Notably, thanks to the tournaments held, new game templates have become popular. In the current season, only 87% of games were played using the most popular template, 2SM4d(3), compared to 98% in the previous season.

*******

The statistics on the dates when games ended are similar. The only difference is that there are fewer drawn-out games, and even more games ending on the 6th day of the week instead of the 7th, which likely indicates a refinement of the statistical results due to a larger sample size and more skillful use of rushing strategies by players.

The longest game this time was between newcomers from the AYMAN portal and Ticcer, and it ended on 3-2-5, which is 4 days shorter than the longest game of the previous season!

The most popular city among players continues to be Castle, probably because it is one of the most neutral cities – not too strong, but not too weak. For the same reason, presumably, the cities considered the strongest and weakest are at the bottom of the list – they are simply excluded or not chosen at the end of the selection process.

This is confirmed below by a graph designed to reflect the strength of the cities. Tower and Inferno are lagging behind. Necropolis and Dungeon are ambiguous cities due to their initial affiliation with underground races. Rampart and Conflux are doing well. And Castle is as always mediocre, neither an ultra-monster nor an outcast. Of course, here we should definitely clarify the strength of the players who chose, say, Tower or Conflux. I mean that if, for example, Pit for LM played all games for Tower, the statistics would be completely different. :D Therefore, this graph cannot be considered a real reflection of the strength of the cities (as well as the heroes below). The data here, in my opinion, is greatly exaggerated. In reality, over a long period (10 seasons, let's say), there would not be such significant jumps on the graph, which would not so clearly indicate which city is obviously stronger or weaker. That is, in fact, Tower is not as weak a city as the graph shows now.

The most popular hero remains Valeska, for the same reasons as Castle. She was chosen in 23% of cases, which means she participated as a starting hero in 45%! of the games. Kayri, Shakti, and Tazar were slightly behind her. And in general, from the graph, we see that for each city, there is one "superhero" that is most often chosen. We don't see any heroes from Conflux (but very few play for it), but there are 2 heroes from Rampart (for which many play).

At the same time, Valeska did not become the best hero again. If you look at the graph, then those heroes will be the "best" whose two graphs come closest to each other. Thus, Tazar has the highest percentage of wins (61). Galtran (59), Kayri (59) and Gurnisson (58) are close behind him. The popular Valeska only won 38% of the games in which she played.

Well, how about it? Beautiful graphs? I agree! But the statistics remained the same (with only minor changes) and therefore practically useless... However, an expert on Heroes III would draw approximate graphs even without all this statistics. And greater accuracy cannot be achieved due to the fact that the strength of the players who played for a particular city is not taken into account in the statistics.

Yes... It turns out that for the next season, we need to provide statistics taking into account the strength of the players. The question is, what coefficients should be used. I think they can be selected empirically. Only again, there is no point in this, because, as I said, "an expert on Heroes III would draw approximate graphs even without..." And the graphs will remain the same as before, only more smoothed.