Discussion in 'Rugby World Cup 2019' started by Umaga's Witness, Sep 2, 2019.
This is a cool idea, fairplay for organising it.
I'm also willing to swap my 11, for any position.
nah it's halfback . but yes.
the original post also has the allocations now
Yes, each of us gets to be the sole judge of our position, to be announced at the end of the tournament. It allows you to have one eye on that player, likely providing a better assessment than what people would see if they tried to rank all positions.
Absolutely will be interesting to post any assessments/thoughts you are having as you go.
17 was just a pretty random number that felt like a reasonable amount of games to watch to qualify as a selector.
are you rubbing his nose in the fact he can't select ringrose?
To be fair it's probably a good thing I didn't get 13, Garry could have gotten injured before the tournament and not played and I'd still have picked him.
But we all know that if he played 9 he'd be the best in the world...
Ok bud, I'm all over it!
14, Sir yes Sir
One never knows maybe a lesser known May, Ioane, Kolbe comes though, but clear favorite is - "Me Skor Tris" - Japan
I will try to suppress the urge to shout “BIG BILL MAAAATTTTAAAAAAAAA” any time I am asked who the best 8 is.
I just resigned and am heading to London at the start of November. The result of this is that I will have nothing to do for most of October so will be able to catch most of the games. Let me know if there are any positions available.
Take mine, loose head. I won't be able to watch the group stages as much as I'd like apart from this weekend really.
respect the sacrifice. Saulan, you are in.
I'm going to try watch every game although I know that there's going to a be few of the T2 teams matches where I probably just up seeing the highlights.
If you are keen to join pick a reserve position and I’ll add you in.
Depending on my work load I may have gracefully bow out - got a couple of deadlines coming up at work which might impact my sneakily-watching-the-games ability
We'll see though
It's ok thanks, I'd rather just criticise everyone else's choices.
I do find it quite hard to pick teams like these, you always tend to gravitate to the T1 teams players and the teams that go furthest. Perhaps I'll try to do the team of the tournament for teams that don't make it to the QF?
Your No.4, you can just put AWJ down now, sorted.
Pfft, all these part time rugby fans
I’ll be recording all games and watching them back just so I can record all the stats I think are relevant for an 8.
If you are serious that would be awesome, you could take the opportunity to invent better stats than are currently in existence. For example, anything that has a subjective component is never a stat partly because different recorders will record different things. But some things, if you are the only recorder, you may be able to achieve a consistent viewpoint over time.
Effective ruck clearance. A clearance providing quick ball.
Effective ruck clearance in second phase after getting over the advantage line. This is score time territory. A lot of tries come down to how well this is done.
Achieving a turnover. Could be by holding someone up, crating a maul; tipping the ball; clean steal; winning a penalty; driving the ruck back over the ball
Contributing to a turnover. When things above are caused by more than one player. Or when a player tackles the opposition behind the advantage line, bringing them to the ground quickly, to allow a teammate to come in and fetch the ball.
You may also want to consider the role the player has in the team, because not every 8 role is the same. Billy v is a ball carrier whereas pocock has a large fetching role, for instance. Point being a player might contribute highly to the team’s effectiveness without looking like the typical 8.
This is exactly what I’m aiming to do. Have broken down the stats into categories of open field, breakdown and set piece for both offence and defence, plus an error category and discipline category. Within each of these categories there are a range of stats (e.g. %of offensive rucks hit which result in ball retained). Loving the suggestions above too on contributing to turnovers!
There’s a few things I’m still not sure about. First one, and this may sound silly, is measuring the decision making element of what 8’s do. So for example where a team concedes a turnover a turnover partly because an 8 chose not to hit the ruck, or a try isn’t scored because he didn’t play the pass. Not sure there’s a way around it so will have a comments column for that kind of thing.
Another thing I’m still considering is the extent to which one applies weighting to different stats. As you mention above different 8s contribute in different ways. I’m hoping with the volume of stats that one could do an 8v8 comparison, but I’m not sure it’ll come out that clear cut.
Last thing that is tbc is do I factor in who the opponent is? So a big stat game against the ABs being “worth” more than a big game against Namibia, or not applying such a weighting? I think it’s goibg to be a case of the games I’m passionate about (NZ and Scotland) I’ll prob watch as a fan then rewatch for analysis, the others I’ll try to do on first watch.
A tricky but fascinating task. think you have to weight depending on opposition and importance but in a way that ensures players from teams that don’t make the quarters still have equal chance. For them the toughest pool games that they have a chance of winning might be their important games.
In terms of weighting individual stats I don’t know. You could try and base it on contribution to territory and possession and potential for points.
1. you could come up with a subjective score for each game as well. These need to be in units such that a score twice as big as another score means the performance was twice as good. So probably not a percentage, because human brains tend to gravitate toward a bell curve distribution when assessing housing percentages (e.g you won’t give a 10/10 very often and, if you do, it doesn’t mean you think the performance was twice as good as one you gave a 5)
2. Normalise your other stats at the end such that the average is 1 for each stat.
3. Optimise the weighting’s to minimise the sum of the squared errors in terms of how well the sum of the stat times the weighting for a given performance predicts your subjective score for that performance (essentially a least squares regression with the weighting’s as the parameters being estimated)
4. Use that prediction as the final score
5. The errors should contain the bulk of any subjective biases that came into the subjective scores.
I might have got a bit geeky there
... would be good to confirm your participation or not so we can try and get a replacement in prior to the first game....
Yeah, I think I'm gonna have to bow out unfortunately - I'm not even sure I'm going to be able to watch the opening game so it's not the best start
All good. It won’t be all that hard to watch all the games from nz. The earliest ones start after 5 pm and the latest ones finish before 1am.
Separate names with a comma.