Monday, September 1, 2014

ETC Match-up Software

The matchup process is critically important and can make the difference between success and failure. I say "can"  because the players still have to 1) have their assessment of the matchup correct and 2) deliver to that assessment.

A couple of years ago Jack wrote software that allowed us to manage the matchups at the 4 person per team NZTC. We've used it for the past two years and I'm happy to say that it adds considerable value. As long as your assessments are right and your team delivers to those assessments then you enhance your results.

The key thing is that it takes the emotion out of the matchup process reducing it to a math-based solution.

So with the ETC coming up I had Jack expand the programme to cater for eight person teams. Having produced the software the key thing was for me to get used to running it and understanding the way it worked.

Critical was ensuring that the input - assessment of matchups - was as solid as it could be. This process took a number of iterations - including a remedial five hour session the night before the event. This was the first time we had had our team all together in the same room.

As noted, the programme works to achieve the best mathematical solution it can. It plays no favourites and therefore there is little emotion in choices.

So how did it work? Well here is a breakdown of the results.

It gives a predicted score when you put up your first choice and then seeks through the process to improve that. The "blue" section is where we ended up and "red" the improvement achieved.

On average, the software enhanced our score by approximately 6 points between Choice #1 and Choice #8.  This is pretty significant in an ETC context when your opponents are trying to do the same. Generally we were left with 4-6 matchups that "promised" 11-13 points, 1-2 8-9 point matchups with the rest 10-10.

From a Captain's point of view it was fantastic. The whole process was straightforward and emotionless. Push the button - and apart from choice #3 that takes 60-90 seconds - instantly have your choice or your preferred opponent in front of you.

Of course it doesn't take in player ability to deliver the result but I always felt we were getting a head start each round. Against Scotland, Australia and Switzerland our performance delivered within 2 points of predicted score - suggesting that our matchups and play was good. In the game against the Czech Republic we were looking good to achieve the same but points slipped away in the last three games.

Our two big losses I think were more down to our play though against Norway at least one of the predicted matchup scores was well wrong.

In the end my thoughts were that our performance could have been improved with more preparedness. The opportunity to play and discuss the matchups to ensure the input into the model was even more "accurate" can't be understated. Our session the night before helped immeasurably. I would have loved two or three more of those informed by tournament play.

However as far as the software goes, it was an unqualified success. It really was our 10th Man.

Thanks to Jack for his time developing it.....and making sure I could drive it.


  1. So you were never tempted to take it in turns to go first? Any chance of this software being uploaded into the App-verse?

  2. Interesting, being on the inside of the Scotland camp, we can point at a totally wrong prediction on our side which screwed our pairings as it got set up and we thought it was in our favour only for the game to go massively to NZ (and our future pairings for that army got reviewed/amended!)

    1. Hi Dave,

      Looking at the pairings sheet we ended up with 5 +ve pairings, two mirrors both of which we liked (DE, Liz) and one -ve matchup (our WoC vs. Skaven). I'm assuming the match you are alluding to was your CDs vs our DoC. You chose that pair and we were ecstatic :-)


  3. Pretty much spot on. Only some of your +ves we had as draws. Would back our DE,Dw and liz players to outperform on avg based on past experience ( all 4+ Years in the team)