Opinion Questions For Rowsus
- Thread starter Rowsus
- Start date
- Status
- Joined
- 12 Apr 2014
- Messages
- 25,401
- Likes
- 65,467
- AFL Club
- Collingwood
Very interesting stuff. Sorry for my slow response, I had a very long day at work yesterday.
My main observation here would be that there are some players that have much higher SC to DT ratios than we (I) might normally think. They seem to be largely defenders, perhaps because they get SC points for less high profile or even “negative” acts like a spoil (preventing their opponent from taking a mark), which DT doesn’t recognize. This seems to be reasonably systematic from what I can see, so perhaps this would need to be adjusted for vs prior years to see any excess scaling in 2020. I’m not sure if this makes things too difficult for you.
A few other thoughts:
- My personal view is that the 2020 scaling effect will be reasonably modest for most players over the course of the season (perhaps c. 5% for many, 10% for a few?). There are a few possible aspects to this as I see it.
- The first is that shortened quarters increase the range of individual players’ scores, because scores are still adjusted so that they total 3,300. To take an extreme example, if a game was reduced to just enough time for one hitout, kick into the forward line and a converted set shot, you might have 3 players getting the full 3,300 points, and everyone else getting zero. The players in this game still average 75, but the distribution is very uneven, and it benefits some enormously and hurts others a reasonable amount. Over a decent sample, this should broadly even out, in that different players will be involved/not involved in these plays. Those who are involved more than you’d expect them to be over a normal season will still have an inflated average though. This will be o***et (I believe perfectly, because all games still have 3,300 points awarded) by those who were involved in those plays less than you’d expect. The magnitude will depend on how big the benefit is, and how often it actually happens versus its expected frequency.
- One practical way to deal with this could be to plot the distribution of individual game scores in 2020 vs say 2019/18 and then adjust 2020 to make it more similar to prior years. This may mean that a 180 in 2020 becomes a 160 from any other year (because, based on frequency, they are similarly rare). Some 20s may become 35s as well, but this probably isn’t a big deal for the players you’re interested in.
- The second aspect is that last season was obviously shortened in term of games as well, which makes the sample smaller. This means that an outsized 180 type score as a bigger impact on a player’s average. The above approach would adjust for this issue fairly straightforwardly I think.
- Injuries also seemed to be up last year, potentially exacerbating the issue with the reduced number of rounds. I think most of the highest averaging players played 80-100% of games though.
- Third, some people have flagged that shorter games allowed higher TOG for certain players and/or a different game style that affected SC scores. Given that this could influence DT scores, as well as all of the underlying stats (kicks, marks etc), it’s not clear to me how this could be adjusted for. You could attempt to TOG adjust, but my view would be that a player’s SC/TOG % would decline as TOG % increases, either because they fatigue, or because they potentially play a less demanding role to help facilitate the change in TOG.
I appreciate this is a bit of detail, but if you are looking to deal with the bulk of the scaling issue without making too many additional assumptions, I think bringing the extreme high end 2020 scores back to what they would have been on a 2018/19 distribution should largely do the trick.
My main observation here would be that there are some players that have much higher SC to DT ratios than we (I) might normally think. They seem to be largely defenders, perhaps because they get SC points for less high profile or even “negative” acts like a spoil (preventing their opponent from taking a mark), which DT doesn’t recognize. This seems to be reasonably systematic from what I can see, so perhaps this would need to be adjusted for vs prior years to see any excess scaling in 2020. I’m not sure if this makes things too difficult for you.
A few other thoughts:
- My personal view is that the 2020 scaling effect will be reasonably modest for most players over the course of the season (perhaps c. 5% for many, 10% for a few?). There are a few possible aspects to this as I see it.
- The first is that shortened quarters increase the range of individual players’ scores, because scores are still adjusted so that they total 3,300. To take an extreme example, if a game was reduced to just enough time for one hitout, kick into the forward line and a converted set shot, you might have 3 players getting the full 3,300 points, and everyone else getting zero. The players in this game still average 75, but the distribution is very uneven, and it benefits some enormously and hurts others a reasonable amount. Over a decent sample, this should broadly even out, in that different players will be involved/not involved in these plays. Those who are involved more than you’d expect them to be over a normal season will still have an inflated average though. This will be o***et (I believe perfectly, because all games still have 3,300 points awarded) by those who were involved in those plays less than you’d expect. The magnitude will depend on how big the benefit is, and how often it actually happens versus its expected frequency.
- One practical way to deal with this could be to plot the distribution of individual game scores in 2020 vs say 2019/18 and then adjust 2020 to make it more similar to prior years. This may mean that a 180 in 2020 becomes a 160 from any other year (because, based on frequency, they are similarly rare). Some 20s may become 35s as well, but this probably isn’t a big deal for the players you’re interested in.
- The second aspect is that last season was obviously shortened in term of games as well, which makes the sample smaller. This means that an outsized 180 type score as a bigger impact on a player’s average. The above approach would adjust for this issue fairly straightforwardly I think.
- Injuries also seemed to be up last year, potentially exacerbating the issue with the reduced number of rounds. I think most of the highest averaging players played 80-100% of games though.
- Third, some people have flagged that shorter games allowed higher TOG for certain players and/or a different game style that affected SC scores. Given that this could influence DT scores, as well as all of the underlying stats (kicks, marks etc), it’s not clear to me how this could be adjusted for. You could attempt to TOG adjust, but my view would be that a player’s SC/TOG % would decline as TOG % increases, either because they fatigue, or because they potentially play a less demanding role to help facilitate the change in TOG.
I appreciate this is a bit of detail, but if you are looking to deal with the bulk of the scaling issue without making too many additional assumptions, I think bringing the extreme high end 2020 scores back to what they would have been on a 2018/19 distribution should largely do the trick.
Hi nico,
sorry it's taken me so long to get back to you. Life getting in the way, and all that.
I think @Darkie has given you some very useful leads in this. With the tables you produced, to compare SC to DT, I would drop out all the players that weren't SC revelant, or just below SC relevant, then look at how it stacks up. As Darkie was very eloquently expressed, the shorter Qtrs, combined with a need to allocate 3300 points, means some of the larger scores get pushed higher, and some of the lower scores get pushed comparatively lower.
As to scaling effect, most of the scaling is actually done "in the run". Someone like @Erich1036 or @Bontempele can probably explain this better, as they have a really good grasp of the make up of a SC score. Here is my understanding of the "in the run" scaling. I will use an overly simple example, just to make it easier for me to explain. The aim is to get each quarter to total roughly 825 points. Now imagine 2 scenarios:
Scenario 1: The game is an untidy tussle, with lots of stoppages, and few clean effective disposals. As we get to, just for example, to 20% of the way through the 1st quarter, there have only been a total of 140 points allocated, including Aspen on 25 points, Brash on 10 points, and Carter on 4 points. (As far as I know, this isn't done at particular points in the quarter, but just done as the quarter progresses.) At 1/5 of the way through the quarter, there should be around 165 points allocated. So everyone's score is bumped by a factor 165/140. So Aspen goes from 25 to 29.5, Brash from 10 to 11.8, and Carter from 4 to 4.7. As we can see, while the scaling was proportional, Aspen was by far the biggest beneficiary.
Scenario 2: Same players, same scores, but it's been open, free flowing, and lots of points allocated. In fact, 20% of the way through the Qtr, there has been 180 points allocated. The same scaling process kicks in, trying to reduce the total to 165 points again. So Aspen goes from 25 to 22.9, Brash goes from 10 to 9.2, and Carter goes from 4 to 3.7.
To the best of my knowledge, this why you can see a players score creeping up, or conversely, going down, when they haven't touched the ball, or aren't even on the ground. What we found last year was, the players that had impact early were getting scaled up a lot in the run of the game. As we saw in the 2 scenarios above, the higher scorers benefit greatly from this kind of scaling. When you factor in the Qtrs were 20% shorter, there was obviously a lot more scaling up going on, than scaling down. I think @Darkie 's estimate of it generally being in the 5% to 10% more than we might normally have seen is probably pretty close, but there will be instances where it is even higher. This also goes a long way to explaining, how players with lowish numbers, that might normally translate to 90 or 95 scores, ended up with much higher scores, as they might have had the score impacting plays early in the quarter, which will possibly have a slightly better nett effect on the score, than doing it late in the quarter. (unless it is late in the last quarter, with the game still on the line.). It can happen that within the one quarter it is going slowly, then suddenly speeds up, as far as allocated points are concerned. In this scenario, it is the "raw" scores that get adjusted in the required direction, and not the "new adjusted scores".
While some players obviously benefitted more than others (Gawn and Neale obviously spring to mind), it is interesting to compare the top CO players from 2019 and 2020.
2019 - The top 5 CP winners averaged 15.9 CP's/game, 28.9 Disps with a DE of 72.1%, SC 113.5, DT 100.9
2020 - The top 5 CP winners averaged 12.5 CP's/game, 23.1 Disps with a DE of 66.0%, SC 118.7, DT 87.0
With shorter Qtrs all their numbers were down, including the DE%, SC went up 4%, but DT went down 14%. All the key scoring areas were down similarly. Marks 3.34 to 3.04, Tackles 4.44 to 4.2, Goals 0.57 to 0.51. This is just looking at those 5 top CP players.
There is no definitive answer, as to how you might adjust their scores. Maybe make a sliding scale, marking all the the 110+ averages down between 7% at the top, sliding to 1% at the 110 mark. It's not easy for you, I know. The other option is just to * the season, and move on.
Good luck, and continue your great work!
sorry it's taken me so long to get back to you. Life getting in the way, and all that.
I think @Darkie has given you some very useful leads in this. With the tables you produced, to compare SC to DT, I would drop out all the players that weren't SC revelant, or just below SC relevant, then look at how it stacks up. As Darkie was very eloquently expressed, the shorter Qtrs, combined with a need to allocate 3300 points, means some of the larger scores get pushed higher, and some of the lower scores get pushed comparatively lower.
As to scaling effect, most of the scaling is actually done "in the run". Someone like @Erich1036 or @Bontempele can probably explain this better, as they have a really good grasp of the make up of a SC score. Here is my understanding of the "in the run" scaling. I will use an overly simple example, just to make it easier for me to explain. The aim is to get each quarter to total roughly 825 points. Now imagine 2 scenarios:
Scenario 1: The game is an untidy tussle, with lots of stoppages, and few clean effective disposals. As we get to, just for example, to 20% of the way through the 1st quarter, there have only been a total of 140 points allocated, including Aspen on 25 points, Brash on 10 points, and Carter on 4 points. (As far as I know, this isn't done at particular points in the quarter, but just done as the quarter progresses.) At 1/5 of the way through the quarter, there should be around 165 points allocated. So everyone's score is bumped by a factor 165/140. So Aspen goes from 25 to 29.5, Brash from 10 to 11.8, and Carter from 4 to 4.7. As we can see, while the scaling was proportional, Aspen was by far the biggest beneficiary.
Scenario 2: Same players, same scores, but it's been open, free flowing, and lots of points allocated. In fact, 20% of the way through the Qtr, there has been 180 points allocated. The same scaling process kicks in, trying to reduce the total to 165 points again. So Aspen goes from 25 to 22.9, Brash goes from 10 to 9.2, and Carter goes from 4 to 3.7.
To the best of my knowledge, this why you can see a players score creeping up, or conversely, going down, when they haven't touched the ball, or aren't even on the ground. What we found last year was, the players that had impact early were getting scaled up a lot in the run of the game. As we saw in the 2 scenarios above, the higher scorers benefit greatly from this kind of scaling. When you factor in the Qtrs were 20% shorter, there was obviously a lot more scaling up going on, than scaling down. I think @Darkie 's estimate of it generally being in the 5% to 10% more than we might normally have seen is probably pretty close, but there will be instances where it is even higher. This also goes a long way to explaining, how players with lowish numbers, that might normally translate to 90 or 95 scores, ended up with much higher scores, as they might have had the score impacting plays early in the quarter, which will possibly have a slightly better nett effect on the score, than doing it late in the quarter. (unless it is late in the last quarter, with the game still on the line.). It can happen that within the one quarter it is going slowly, then suddenly speeds up, as far as allocated points are concerned. In this scenario, it is the "raw" scores that get adjusted in the required direction, and not the "new adjusted scores".
While some players obviously benefitted more than others (Gawn and Neale obviously spring to mind), it is interesting to compare the top CO players from 2019 and 2020.
2019 - The top 5 CP winners averaged 15.9 CP's/game, 28.9 Disps with a DE of 72.1%, SC 113.5, DT 100.9
2020 - The top 5 CP winners averaged 12.5 CP's/game, 23.1 Disps with a DE of 66.0%, SC 118.7, DT 87.0
With shorter Qtrs all their numbers were down, including the DE%, SC went up 4%, but DT went down 14%. All the key scoring areas were down similarly. Marks 3.34 to 3.04, Tackles 4.44 to 4.2, Goals 0.57 to 0.51. This is just looking at those 5 top CP players.
There is no definitive answer, as to how you might adjust their scores. Maybe make a sliding scale, marking all the the 110+ averages down between 7% at the top, sliding to 1% at the 110 mark. It's not easy for you, I know. The other option is just to * the season, and move on.
Good luck, and continue your great work!
- Joined
- 6 Jul 2012
- Messages
- 2,301
- Likes
- 1,721
- AFL Club
- North Melb.
Thanks for the shout out Rowsus! Nicohighscore - without going too much into the detail here, I think I probably wouldn't worry too much about inflated scoring last year. The games were 64 mins instead of 80 mins, so basically everyone should've gotten 80% of their usual stats, but gotten scaled up to the usual 100%.
Unless certain players were particularly good in shortened game time, maybe NicNat for example, it shouldn't make too much difference. Perhaps if you could identify players that had spikes in % time on field last year, you might want to stay away from them.
Some have noted that the highest scoring players were exceptionally high in 2020. Again, I wouldn't worry too much about this, because stoppages were at record-high numbers, which favours the top mids and rucks.
Unless certain players were particularly good in shortened game time, maybe NicNat for example, it shouldn't make too much difference. Perhaps if you could identify players that had spikes in % time on field last year, you might want to stay away from them.
Some have noted that the highest scoring players were exceptionally high in 2020. Again, I wouldn't worry too much about this, because stoppages were at record-high numbers, which favours the top mids and rucks.
The big question on that though is were the shorter quarters the cause of those stoppage numbers and will that revert backwards and/or will rule changes also impact this?
I agree we can overthink it a bit too much, ultimately if all the premiums drop 5% then it's only any additional drop that you really need to worry about!
I agree we can overthink it a bit too much, ultimately if all the premiums drop 5% then it's only any additional drop that you really need to worry about!
- Joined
- 1 Nov 2019
- Messages
- 435
- Likes
- 2,565
- AFL Club
- Richmond
Thanks guys for your valuable comments.
There is another question I keep bumping on - but I reckon this is SC 101 and it must have been discussed extensively in the past, and I am just too novice to not have a clear view about it.
I remember last year spending all this time to have the perfect starting team, especially with all the best rookie/cash cow possibles - which I kinda nailed. Except that very quickly, I realised that not having any loophole option (because no donut ) was bitting me hard...
So - what is the balance between going for 1 perma donut on each line, (with DPP if you have $ to spend?) - to have the ability to loophole 1 player per line and ensure a VC / C choice ,
and,
gambling on rookie, with the potential cash it produces (over time...) but the loss of point the above can generate.
There is another question I keep bumping on - but I reckon this is SC 101 and it must have been discussed extensively in the past, and I am just too novice to not have a clear view about it.
I remember last year spending all this time to have the perfect starting team, especially with all the best rookie/cash cow possibles - which I kinda nailed. Except that very quickly, I realised that not having any loophole option (because no donut ) was bitting me hard...
So - what is the balance between going for 1 perma donut on each line, (with DPP if you have $ to spend?) - to have the ability to loophole 1 player per line and ensure a VC / C choice ,
and,
gambling on rookie, with the potential cash it produces (over time...) but the loss of point the above can generate.
Thanks guys for your valuable comments.
There is another question I keep bumping on - but I reckon this is SC 101 and it must have been discussed extensively in the past, and I am just too novice to not have a clear view about it.
I remember last year spending all this time to have the perfect starting team, especially with all the best rookie/cash cow possibles - which I kinda nailed. Except that very quickly, I realised that not having any loophole option (because no donut ) was bitting me hard...
So - what is the balance between going for 1 perma donut on each line, (with DPP if you have $ to spend?) - to have the ability to loophole 1 player per line and ensure a VC / C choice ,
and,
gambling on rookie, with the potential cash it produces (over time...) but the loss of point the above can generate.
There is another question I keep bumping on - but I reckon this is SC 101 and it must have been discussed extensively in the past, and I am just too novice to not have a clear view about it.
I remember last year spending all this time to have the perfect starting team, especially with all the best rookie/cash cow possibles - which I kinda nailed. Except that very quickly, I realised that not having any loophole option (because no donut ) was bitting me hard...
So - what is the balance between going for 1 perma donut on each line, (with DPP if you have $ to spend?) - to have the ability to loophole 1 player per line and ensure a VC / C choice ,
and,
gambling on rookie, with the potential cash it produces (over time...) but the loss of point the above can generate.
Hey Slams,
to me Gawn is sort of killing two birds with one stone.
As you are familiar with, my road map for picking your team is pick your Captain options first (no matter what the price!), then decide your Ruck strategy as your next step. To me, Gawn fills both those steps, so you're sort of getting two picks in your selection process, in the one pick!
As for McEvoy in the backline, keep in mind, he's not much use as a Ruck replacement if there are late outs. To get the full benefit of McEvoy as Ruck cover, you need to structure up something like this:
Def: McEvoy
Ruck: F/R
Fwd: D/F
Unfortunately, it is not as easy as rotating them around, as it is a "triple rotation". The only way to move them to where you need them, if you need McEvoy on your Ruck line is by having a trade in hand.
ie. Trade out your D/F, swap your F/R down to the Fwd line, swap McEvoy down to Ruck, trade the same D/F player back into your Def line. That way you have achieved the position swaps, without burning a trade.
I think Grundy might have been a victim of circumstances, and the tight Draw last season. My thinking is, he picked up a niggle, and because of short back ups, and Collingwoods need to win games, to make the finals, he soldiered on at less than 100%.
If I can get a playing R3, I'm likely to start: Gawn, Martin // R3.
If I can't get a playing R3, I might just go Gawndy.
to me Gawn is sort of killing two birds with one stone.
As you are familiar with, my road map for picking your team is pick your Captain options first (no matter what the price!), then decide your Ruck strategy as your next step. To me, Gawn fills both those steps, so you're sort of getting two picks in your selection process, in the one pick!
As for McEvoy in the backline, keep in mind, he's not much use as a Ruck replacement if there are late outs. To get the full benefit of McEvoy as Ruck cover, you need to structure up something like this:
Def: McEvoy
Ruck: F/R
Fwd: D/F
Unfortunately, it is not as easy as rotating them around, as it is a "triple rotation". The only way to move them to where you need them, if you need McEvoy on your Ruck line is by having a trade in hand.
ie. Trade out your D/F, swap your F/R down to the Fwd line, swap McEvoy down to Ruck, trade the same D/F player back into your Def line. That way you have achieved the position swaps, without burning a trade.
I think Grundy might have been a victim of circumstances, and the tight Draw last season. My thinking is, he picked up a niggle, and because of short back ups, and Collingwoods need to win games, to make the finals, he soldiered on at less than 100%.
If I can get a playing R3, I'm likely to start: Gawn, Martin // R3.
If I can't get a playing R3, I might just go Gawndy.
Thanks for the good words.
I think you can find most of the answer to your question at posts #10,127 & #10,138.
I'm not sure Neale and/or Gawn dropping $100k in price is the problem people are making it out to be. The bottom line is, you just need to maximise the points you are getting, from the $7,000,000 to $7,600,000 you spent on players, with the intention of holding them through the season, if possible.
It's a bit simplistic to say, Coach A had Gawn from the start. Coach B picks him up a bit later at $100,000 less. It really depends on how well Coach B has used the money he didn't spend on Gawn.
To try and put a mathematical sum to your question. Let's assume a cut off point at Round 8, and the value of a point by that stage (ie. the Magic Number) is around $5,100. So $100,000 is worth around $100,000 / $5,100 = 19.6 points.
The question then becomes, do you get the benefit from that $100,000 saved on Gawn
in the first 8 Rounds then 8 x 19.6 = 157 points
the 15 Rounds after that = 15 x 19.6 = 294 points
The whole 22 Rounds = 22 x 19.6 = 431 points.
I think the best way to look at it, isn't "I could save $100,000 if I wait", but to say, "What do I think Gawn/Neale will score in those first 8 Rounds, and do I think I can pick Captain options that are their equal, or better, for less cash?".
Answer those 2 questions, and you'll have the answer as to whether it is worth starting those expensive players.
I think you can find most of the answer to your question at posts #10,127 & #10,138.
I'm not sure Neale and/or Gawn dropping $100k in price is the problem people are making it out to be. The bottom line is, you just need to maximise the points you are getting, from the $7,000,000 to $7,600,000 you spent on players, with the intention of holding them through the season, if possible.
It's a bit simplistic to say, Coach A had Gawn from the start. Coach B picks him up a bit later at $100,000 less. It really depends on how well Coach B has used the money he didn't spend on Gawn.
To try and put a mathematical sum to your question. Let's assume a cut off point at Round 8, and the value of a point by that stage (ie. the Magic Number) is around $5,100. So $100,000 is worth around $100,000 / $5,100 = 19.6 points.
The question then becomes, do you get the benefit from that $100,000 saved on Gawn
in the first 8 Rounds then 8 x 19.6 = 157 points
the 15 Rounds after that = 15 x 19.6 = 294 points
The whole 22 Rounds = 22 x 19.6 = 431 points.
I think the best way to look at it, isn't "I could save $100,000 if I wait", but to say, "What do I think Gawn/Neale will score in those first 8 Rounds, and do I think I can pick Captain options that are their equal, or better, for less cash?".
Answer those 2 questions, and you'll have the answer as to whether it is worth starting those expensive players.
Thanks guys for your valuable comments.
There is another question I keep bumping on - but I reckon this is SC 101 and it must have been discussed extensively in the past, and I am just too novice to not have a clear view about it.
I remember last year spending all this time to have the perfect starting team, especially with all the best rookie/cash cow possibles - which I kinda nailed. Except that very quickly, I realised that not having any loophole option (because no donut ) was bitting me hard...
So - what is the balance between going for 1 perma donut on each line, (with DPP if you have $ to spend?) - to have the ability to loophole 1 player per line and ensure a VC / C choice ,
and,
gambling on rookie, with the potential cash it produces (over time...) but the loss of point the above can generate.
There is another question I keep bumping on - but I reckon this is SC 101 and it must have been discussed extensively in the past, and I am just too novice to not have a clear view about it.
I remember last year spending all this time to have the perfect starting team, especially with all the best rookie/cash cow possibles - which I kinda nailed. Except that very quickly, I realised that not having any loophole option (because no donut ) was bitting me hard...
So - what is the balance between going for 1 perma donut on each line, (with DPP if you have $ to spend?) - to have the ability to loophole 1 player per line and ensure a VC / C choice ,
and,
gambling on rookie, with the potential cash it produces (over time...) but the loss of point the above can generate.
Let's assume you start 14 non-Rookies (includes Mid Price players) and 16 Rookies.
You are trying to build a successful list, of let's say 12 of the better Rookies.
If you have no Donuts, there are 16x15x14x13/4/3/2/1 = 1,820 combinations of 12 your 16 Rookies make, giving you a lot of chances of achieving your quest for 12 of the better Rookies.
If you have 3 Donuts you now have 13 Rookies, to try and find your 12 better Rookies. That only gives you 13 combinations of 12.
You are 140 times more likely to hit the 12 Rookies you want/need, with no Donuts, than you are with 3!!!
The other thing to keep in mind this season, is that there is no Thursday night team announcements. This means you want as much playing stock at your finger tips as possible, because you won't know on Friday, or possibly even on Saturday, who is in the 22 on Sunday!!!
My best advice this season, is aim for 30 starters in Round 1, or at least 29 if you need the Captain's loophole in Round 1!
There is Bigoa Nyuon at 123k who has R/D as well, so you can do it without the triple, costs you 22k on Treacy obviously though.
Still, McEvoy needs to be a good pick on his own, imo, to really start looking at that path.
You also have to be good (lucky) at loopholing. I personally would prefer the cash over a donut in the first round, reality is that most seasons you've got a donut by round 3 anyway so why start them.
But I'm terrible at loopholing, at one point I worked it out last year that I'd cost myself over 1000 points from loopholes and I was even worse the year before (admittedly where I was overseas and thus missed a couple). I even copped the dreaded "late in" captain, can't remember who it was but it hurt.
Still, McEvoy needs to be a good pick on his own, imo, to really start looking at that path.
You also have to be good (lucky) at loopholing. I personally would prefer the cash over a donut in the first round, reality is that most seasons you've got a donut by round 3 anyway so why start them.
But I'm terrible at loopholing, at one point I worked it out last year that I'd cost myself over 1000 points from loopholes and I was even worse the year before (admittedly where I was overseas and thus missed a couple). I even copped the dreaded "late in" captain, can't remember who it was but it hurt.
- Joined
- 1 Nov 2019
- Messages
- 435
- Likes
- 2,565
- AFL Club
- Richmond
Just to close off this topic - I have tried to (pseudo) mathematically have an estimate of :
I have a VC + C option available each round, COMPARED TO I only go direct for the C.
@Rowsus - I am not very confident in my statistical approach ... feels like a bit of a /roastme post ...
Taking VC Oliver, mean of 123, Std dev of 30 and Gawn, Mean of 140 and Std Dev of 30.
Considering Normal distribution.
I get a probability for each score range for both players :
From this point, I can compare the distribution of my Captain score , if I go either direct to Gawn or if I go VC Oliver first, and keep his score only if it is over 120.
I get a score distribution for my Captain score in both case :
Conclusion : in this case, you don't necessary gain much in average ( 5 pts X 2 ) - but you divide by 2 your probability to have a C score below 120 - which can be a huge difference if you want consistency for league matchs... ?
If you use Neale and Then Gawn - the result is even better - as expected :
7 pts x 2 over in average, you almost divide by 3 your probability of a disappointing C score
I guess I can see how in terms of pure value, it isn't worth the 450k 3 good rookies could bring you in about 5-6 rounds...
(plus the security of having active players on the bench for last minute change & injuries)
I have a VC + C option available each round, COMPARED TO I only go direct for the C.
@Rowsus - I am not very confident in my statistical approach ... feels like a bit of a /roastme post ...
Taking VC Oliver, mean of 123, Std dev of 30 and Gawn, Mean of 140 and Std Dev of 30.
Considering Normal distribution.
I get a probability for each score range for both players :
From this point, I can compare the distribution of my Captain score , if I go either direct to Gawn or if I go VC Oliver first, and keep his score only if it is over 120.
I get a score distribution for my Captain score in both case :
Conclusion : in this case, you don't necessary gain much in average ( 5 pts X 2 ) - but you divide by 2 your probability to have a C score below 120 - which can be a huge difference if you want consistency for league matchs... ?
If you use Neale and Then Gawn - the result is even better - as expected :
7 pts x 2 over in average, you almost divide by 3 your probability of a disappointing C score
I guess I can see how in terms of pure value, it isn't worth the 450k 3 good rookies could bring you in about 5-6 rounds...
(plus the security of having active players on the bench for last minute change & injuries)
Attachments
Hey BIGBOY,
my apologies, I'm not sure how I missed/skipped over this one.
Butters $471,400 - Fwd - my first thoughts on Butters were, that it might be one season too early for him (He's not even 21 yet!). He hasn't put on much bulk, and might be a bit light in the contest. Against that, he averaged 87.7 last season, and is still available as a Fwd this season, so he would appear to have some upside! It's hard to see him going backwards too much, but you probably want him getting close to 100 if you take him. Given there are so few Fwds we can be confident in, he might be worth the punt. Not really on my radar right now, I must admit.
Simpkin $499,700 - Mid - back in post #10,100, when talking about LDU, I wrote:
........ does North have the game to be able support a 110 Mid? Let's look at recent history:
2020: Anderson 15/105, Dumont 17/101, Higgins 17/95, Simpkin 17/93
2019: Cunnington 22/102, Higgins 17/95
2018: Higgins 20/103, Cunnington 22/96
That's every North Mid that has played more than 10 games, and averaged 90+ in the past 3 seasons. It's not very inspiring. Even with the gap left by Higgins leaving, LDU is a player I'm really hoping to avoid. When push comes to shove, if he has played out of his skull in the pre-season, I will definitely be tempted. I'm just hoping to resist that temptation, as he looks like a scoring or game count trap, or both!
I still think I'd be a bit hesitant to take a Keeper Mid from North.
my apologies, I'm not sure how I missed/skipped over this one.
Butters $471,400 - Fwd - my first thoughts on Butters were, that it might be one season too early for him (He's not even 21 yet!). He hasn't put on much bulk, and might be a bit light in the contest. Against that, he averaged 87.7 last season, and is still available as a Fwd this season, so he would appear to have some upside! It's hard to see him going backwards too much, but you probably want him getting close to 100 if you take him. Given there are so few Fwds we can be confident in, he might be worth the punt. Not really on my radar right now, I must admit.
Simpkin $499,700 - Mid - back in post #10,100, when talking about LDU, I wrote:
........ does North have the game to be able support a 110 Mid? Let's look at recent history:
2020: Anderson 15/105, Dumont 17/101, Higgins 17/95, Simpkin 17/93
2019: Cunnington 22/102, Higgins 17/95
2018: Higgins 20/103, Cunnington 22/96
That's every North Mid that has played more than 10 games, and averaged 90+ in the past 3 seasons. It's not very inspiring. Even with the gap left by Higgins leaving, LDU is a player I'm really hoping to avoid. When push comes to shove, if he has played out of his skull in the pre-season, I will definitely be tempted. I'm just hoping to resist that temptation, as he looks like a scoring or game count trap, or both!
I still think I'd be a bit hesitant to take a Keeper Mid from North.
- Status