# TournamentWorld Cup of Randbats 2023 - Qualifying Round

Status
Not open for further replies.

UUPL Champion

#### Dark Pulse

Now that all my games are done, I can focus on what really matters: the stats

I've been gathering my own independant stats for the qualifiers, while also running some simple predictive models to estimate how it would conclude

First and foremost, I'd like to congratulate teams US South and France for already having their place in the tournament guaranteed - USS had achieved this in the first week, but France just managed to get over the line while I was asleep. Even if they lost all their remaining rounds, and every other team won all theirs (an impossibility of course, but regardless) USS would end the qualifiers in 1st place (!) while France would end in 5th. A great showing from both of you!

Next, while most teams are nearing the end of their games, The tournament is far from over. Bangladesh are having a surprisingly good run, and are currently in a the running to knock UK out of the qualifiers! There are still more games to go, however, and anything could happen.

Finally, I'll explain the 4 columns on the right. These are my simple predictive models. They use the same formula, however are seeded with different numbers.
"Winrate" is seeded by the Game winrate of each player, and cross-checks that with the winrate of each opponent that has not yet been played. Fractional points are then awarded based on the probability of victory. For example, Peap and Trade have not yet played against each other. Peap has a winrate of 71.4% so far in this tournament, while trade's winrate is 50% on the nose. Without explaining the math, this means that the prediction would award 0.714 points to peap, and 0.287 points to trade.
"GxE" instead uses, well, the GxE of players. Unfortunately a bunch of people here don't play ladder, so this data ended up being quite innacurate. For the main event I might try and scrape previous tournament results to get further data on players.

The last 2 columns are repeats of the first 2, however they do not take into account completed games, and pretend we restarted this tournament from the first game. This helps highlight the inconsistancies of using GxE as a measurement.

We may be nearing the end of the qualifiers, but there's still plenty more games to play, and so plenty of more chances for these numbers to change. I can't wait to see where everything lines up, and how accurate my models end up looking

(ps, I first started making this because I didn't realise that the official spreadsheet had a "team standings" tab, so... stupidity is the mother of invention?)

#### Irpachuza

##### You didn't get this far by giving up, did you?
All of this is very cool, great job!

For the main event I might try and scrape previous tournament results to get further data on players.
Feel free to grab the info of all RBTT's from the link on my signature (also available on our website HoF)

#### Stallion

##### Tree Young
Calling activity vs Angel tiopo

It's 35 mins after our scheduled time (30 mins after Drifting contacted Team Mexico) and he or a sub has not shown up.

Furthermore, he hasn't been online since Tuesday Australian time, so yeah.

I don't like calling activity, but I don't like people not respecting my time enough to tell me they can't make it, so...

Cc: sharpclaw

Last edited:

#### Wigglytuff

##### mad @ redacted in redacted
"Winrate" is seeded by the Game winrate of each player, and cross-checks that with the winrate of each opponent that has not yet been played. Fractional points are then awarded based on the probability of victory. For example, Peap and Trade have not yet played against each other. Peap has a winrate of 71.4% so far in this tournament, while trade's winrate is 50% on the nose. Without explaining the math, this means that the prediction would award 0.714 points to peap, and 0.287 points to trade.
how are these calculated? 0.714 seems like 71.4/100 but 0.287 seems to come out of nowhere other than possibly being 1 - 0.714 with a rounding error. is this something similar to strength of schedule / what are its advantages over strength of schedule?

The last 2 columns are repeats of the first 2, however they do not take into account completed games, and pretend we restarted this tournament from the first game.
if "WR" doesn't take into account completed games and restarts from 0, what are the above fractional points that are then used to calculate "WR" derived from?

"GxE" instead uses, well, the GxE of players. Unfortunately a bunch of people here don't play ladder, so this data ended up being quite innacurate.
i agree that it'd be difficult to find gxe's of players, but how did you end up with sub 40 gxe's for any team? it takes a concerted effort of forfeiting every game to hit below 50, let alone the 11 mexico has in your stats

Status
Not open for further replies.