A little history

There is an ongoing debate in the Kubb world, and it’s a monster with two heads: First, “How much advantage is there (if any) in throwing first?” Secondly, and perhaps more importantly, “Even if there IS an advantage, is that a bad thing?” (Or maybe more accurately – “How much advantage is too much”?) I hear the debate at tournaments, on social media, and in my own backyard. There are many different thoughts, ideas, and opinions on the matter, from your casual backyard player to the competitive powerhouses vying for a championship.

 

Before we go on, you should know that a lot of research and discourse has already been poured into this topic. I would assume that most people reading this blog are up on the discussion, but if not, please take some time & catch up – trust me, this post will still be here when you get back.

 

Start here: Basel-3 Rule Change Eliminates Head-Start – Our friends in Switzerland do an excellent job outlining the issue, and even provide a smart little simulation tool you can play with to get an idea of the probabilities they are discussing. Stick around for the comments on the post as well, because there is some excellent discussion generated.

 

This Ask Planet Kubb thread is linked in the comments above, but I’ll include it here as well. You might spend a little more time poking around that site because this certainly isn’t the only time the T1 advantage has been discussed there, but (at least so far) it has the prettiest charts.

 

I was recently on Fox Valley Kubb Radio discussing the matter (as well as some possible solutions) in some detail.

 

In the simplest terms the argument is this: High performing teams, whether throwing first (Team A) or second (Team B) will advance the game quickly towards a resolution. The earlier a game is ended, the more likely it is that a high percentage of the total batons thrown will have been thrown by Team A. The inference is that if one team has more batons (and therefore more opportunities to advance the game), then that team has more opportunities to win. Given equally performing teams, the team with more opportunities to win will win more often.

 

Based on this premise, the theoretical answer to whether or not an advantage exists seems to be “Yes, but it has a greater impact in short games where the teams are performing well. The longer a game lasts, the less of an effect throwing first will have.”

Testing the theory

Ideally, we could look at the results of thousands (or better still, millions) of actual recorded games and analyze the results. Unfortunately, we have a dearth of data surrounding this. We do have at the time of this writing, roughly 3500 matches recorded, but nothing about the individual games, such as which team threw first in each, how many rounds they lasted, etc. We DO have a few more than 100 games recorded throw-by-throw, but that is still a pretty small sample set to try to draw game-level conclusions from.

 

I was intrigued by the simulator the Swiss put together, and I tried my hand at making my own. The three key performance indicators I used to describe the teams being simulated were: 8m accuracy, ‘other than 8m’ accuracy, and Kubbs per Hit in the field – i.e. when field kubbs are hit, how many kubbs, on average, are toppled?

 

Using the data gathered by Planet Kubb (along with some additional kubb videos I’ve seen online), I’ve pieced together by-baton statistics that I am confident in. We have thousands of throws at all distances (over 5,000 so far), and I can say with a great deal of confidence that over the past few years, competitive-level Kubb has seen baseline-to-baseline accuracy rates of about 35% and about 80% accuracy at everything else. When field kubbs are hit, about 1.6 field kubbs are toppled on average.

 

I used these stats to describe my two simulated teams, and let them battle it out virtual-style. By design, the model disregards all other variables. On this ideal pitch there is no wind, the pitch is perfectly level, there are no nerves to increase or decrease performance, no anxiety about being behind or confidence from playing with the lead, there is no “1st baton slump/2nd baton bump”, and finally – most importantly – the teams have an identical potential for performance. In this environment, where the only variable is throwing order, it can be shown that there is a significant correlation between throwing first and winning the game – using the above averages as a benchmark, Team A wins roughly 65% of the time (almost twice as often as Team B) and the trend only increases as performance levels rise.

 

So if the model predicts a 65/35 split, then why doesn’t the actual recorded data (not to mention personal experience) match that? Current data suggests about a 55/45 split, but remember what I said about a small sample set of game level data; the margin of error for a sample this small is about 7%, which DOES encompass a 50/50 split (which would translate as no measurable advantage at all), and DOESN’T encompass the 65/35 from the model (so even if all of the data recorded is aberrantly low, it is incredibly unlikely that Team A’s win average would ever climb to 65% given current performance levels).

 

Secondly, if the advantage conferred is so significant, why don’t we see more matches go to a 3rd game? The model predicts that roughly 54% of matches should be decided in game three, but thus far we have only observed that about 30% make it past game two. So what gives?

 

Well, in a word; performance. Let me be clear here – as much as I decry the imbalance of the 6-6 open, the difference in performance is BY FAR the single largest factor in the outcome of a game. Taken in aggregate, the performances of Team A and Team B are very similar (and that makes sense as every competitive team plays both roles at least once per match so everybody is contributing to both numbers), but the performances in any particular game are far from identical. In the blow-by-blow games we have, the winning team tends, on average, to outperform the other team by around 40 percent.

 

 

Check back tomorrow for Kubb Economics 101, and how we came to that ‘40%’ figure…