Month: April 2017

Complete Information Season 2 Stats Blog

Are you sitting comfortably? Then I’ll begin.

Originally intended to be a brief summary of the season this has become something of a magnum opus in the eyes of those around me as more and more data has been pulled, collated, corrected and finally analysed. As such, this may slightly exceed the intended length and ramble on, I hope only to keep it interesting throughout and present my take on the brilliant season through the lens of statistics.


The Only Points That Matter: Final Standings

Of course the only numbers that matter at the end of the day are league points, they decide our winners and track our season. Winning the Major League with 23 of a potential 28 points is The Other Team. Meanwhile Exodunks take the Minor League with 21 points.

The match wins that can be seen here effectively represent how the scores would line up without bonus points for dominations and show how, especially in the Major League, these bonus points are key in differentiating an otherwise very tight field. Though eventually not an issue the domination wins of Team Johnny and Clevergirls would have left them tied for 4th to be split by Pf. Comparatively V0LTA’s pair of domination victories put them above Provi-Shotz and earnt them 10 additional prize skins.

Overall, despite the outliers in 1st and 8th, the Major League was insanely tight with only 6 points between 2nd and 7th. The Minor League was a more diverse field, but even they found 3rd and 4th divided by but a single bonus point. In total 18 bonus domination points were awarded throughout the season; enough to place 2nd in either league.

Pf(Points for) against Pa(Points against) makes a rather satisfying graph. The S shaped curve follows what you’d expect and bar one team, which I’ll explore later, there are few deviants. In fact it was so consistent I went on to compare the actual Pfs to predicted Pfs based on the average Pf in wins and losses and a team’s record.

Seen here the predictions hold up fairly well, barring an aforementioned selection. Though this is potentially a product of a small data pool it also points to a consistent pattern of break points in matches that is simply a facet of the format.


Red vs Blue: Theorycrafting Really Is All That

One of the hot button topics throughout the season has been the Red vs Blue balance. Once more Bei gets to sit back and relax at the end of season as it miraculously comes out at a perfect 50:50. This seemed a distant thought when we reached the mid season with nearly a 70% win rate for Blue. However this has been turned on its head, especially for the major league where the win rate for Red has been over 85% in the second half of the season and the final day lead us to a full Red wash.

The onus for this shift has been in theorycrafting, Blue having the key first picks of Cruiser, that is to say logistics, and Battlecruiser allowed them to secure composition defining picks. Where Frigate and Destroyer were relatively uncontested groups Battleship could be controlled by the choice of logistics paired with the Blue side split double logistics ban we came to see so frequently. 26 out of 56 matches in fact, and 4 more where a double logistics ban hit just shield or armour rather than the more typical both.

This gave Blue the advantage as the format was felt out, able to find the key picks they needed to put together strong compositions and able to bully Red into potential suboptimal frigate picks left them in a strong position. This shifted around the midpoint of the Championship, though signs were there into Q2.

A number of strategies came together for Red side including: a shift in bans away from an initial theme of armour battleships, the use of frigate logi, griffin picks or afterburner fit frigates all contributed to mitigating or undoing the effects of the Blue side double logistics ban.

This combined with further testing and theorycrafting based around alternative picks, what Red could do after Blue had taken the tools they wanted began to come together to restore and even sway the balance between Red and Blue. With Red still feared when the coin flip comes around it’s been the focus of theorycrafting and swung round to become the dominant side, at least in the Majors.

The drafting differences between the Majors and Minors are quite startling with Blue side priority, and by extension Red side tendency, being almost mirrored. We have also seen more typical drafts from Red in the Majors, seeking logistics and playing out more normal games.

This divide between Majors and Minors in Red vs Blue winrate follows this idea that theorycrafting became the driving factor.  It’s not a difficult leap to take to say that our Major League teams put more in and are potentially just better theorycrafters. This is something that may become even more true next season as we see for the first time in this format a fully qualified Major League.


Armour vs Shield: A Question as Old as Ravens

So, the end of the season has come and we have an answer, armour is better, with over a 60% winrate. By comparison shield only manages a little under 45% bested even by the, admittedly shield based frigate logi picks and their 50%. Not picking Logi however has been, perhaps unsurprisingly, unsuccessful.

This can in part be traced back to the Red and Blue side priorities and win rates for the different leagues. The Blue dominated Minors prioritise armour on Blue, while the more Red dominated Majors prefer to pick Shield on Blue and leave Red to often pick armour. The Armour-Shield divide is most apparent in the Minor League however.

Compared to relatively even numbers in the Majors, Armour has a drastically higher win rate in the Minors and Shield is no longer breaking even. Surprisingly logistics frigates manage an even 50% in both Leagues, though typically the frigates were different, where the Bantam was the exclusive choice in the Major League the Minors saw the Burst in favour despite its worse record.

Interestingly the strategy of playing without logistics is more popular, and far more successful in the Major League. The sample is small enough that this statistic is artificially inflated by Phalanx’s hail mary compositions under Grookshank in the final weekend. However the additional success points to more practised and polished strategies, even for all in no logistics compositions.

Continue reading “Complete Information Season 2 Stats Blog”