Hi Rowsus,
Great work again. I just love your detailed analyses.
Do you have similar lists for the other positions, or are they too speculative to have much value?
Hi Bridge, thanks for the kind words.
There's no doubt the lists are speculative, and based upon my preason predictions.
I don't need to assemble the lists as such, it's an on going function of the full system of
RAMP. That means I have it for all positions, and of course, it is based on my preseason guestimates as to the players performances this season. Anyone that has my version of
RAMP that I sent out preseason can do similar lists, they just need to make sure they keep the scores updated week by week, and know how to write a simple formula and use the sort function. (ok, now I've written that out, I can see some people might struggle to do that.)
Also I noticed the (M8)M10-M15 are very close together (24/71 points). This makes me suspect that their are many players within 100 PIT points of M15, I would hazard a guess that there are probably 10-15. Is that correct?
It also feels that these differences are so small and the amount of guesswork that went into compiling the list in the first place so great that they have hardly any statistical relevance. So many points hang on a little bit of luck in either direction (the tagger went to the team mate, the coach used the player in an unusual (lower scoring) role, a niggling injury, the rookie that filled the spot had a great/disaster week, champion data (or their data entry people) adjusted a players score exceedingly, the player kicked the winning goal after the siren, etc.
Great guess, I just checked, and there are 16 players within 100 points of my M15!
And yes, there is no statistical relevance to the tables. It is all based on my preseason opinion, and what
RAMP predicts will happen from here. There are, as you listed, so many variables that can impact a players scoring, in both directions. All we can do, when trying to predict the future, is to study the past, and see how that information helps us to predict what might happen. It's by no means perfect, but it's as good a system as any.
I was surprised with a few names on your list and would love to hear a little bit more about your reasoning:
I expected D Beams to jump back to his 2012 form, which would see him in the top few. Apart from coming back to the mean, what made you rate him lower.
It appears that you expect Liberatore to increase his season average on last year. What is your basis for this?
You seem to believe that Cotchin will bounce back close to his 2012 average. Why?
It looks to me that you have lowered Selwood's score. How many games do you expect him to miss during the season?
I had:
D Beams marked at 19 games @ 110.9/game coming into the season. His history shows he is likely to miss games at some stage, and I can't see him bouncing straight back to his 2012 numbers. That had him scoring 2440 if he played all 22 games. The difference between an 80 point Rookie, and
his expected average from here is 29.45. So we get his guestimated PIT80 score for the remainder of the season as:
2440 - 251 points already scored - 3 x 29.45 for Rookie replacement score difference = 2,101
Liberatore finished last season so strongly, I was backing him to improve further this season, even though my own research and articles suggested he was really needing to put himself in the top bracket to achieve that. He averaged 117.5 in the last 13 games last season, and that included some good scores against good teams (Haw 120, Ess 125, Syd 118). I marked him at 21 games @ 111.4 for 2014. To achieve that from here:
2450 - 190 points already scored - 1 x 33.00 for Rookie replacement score difference = 2,227
Cotchin was touted as resting Forward more this season. I was hoping this would lead to Selwood type performances, where he might kick 25-30 goals for the season (still might happen), and boost his score accordingly. I had him pegged at 22 games @ 113.6 coming into the season. To achieve that from here:
2500 - 202 points already scored - no adjustment for expected missed games = 2,298
J Selwood I consider to be the 4th true "Super Premium", but he can find his problems. Pre season I had him down for 20 games @ 117.7. That leads to from here:
2590 - 297 points already scored - 2 x 34.65 Rookie replacement score difference = 2,224
It might seem strange that I have Cotchin and Libba down to be worth more points-wise from here than Selwood, but I'm a big believer in things evening themselves out, and history being hard to create. It is so easy to say "Gee, I thought Selwood would be a 118 player this season, but the way he has started, I think he'll be a 129 player now!". That sort of knee jerk reassessing is what leads to poor trading decisions during the season. Birchall last season is a wonderful example of knee-jerk reassessments. Most would have pegged him as a 95/game player coming into the season. Suddenly he shoots off 122, 126, 110 and 152, and everyone is tripping over themselves in excitement! "He's a "must have", and the sooner the better! Damn the price, he will be a 110 player this year "for sure"! (Zorko this year, anyone?). Not only did the inevitable correction come, he actually finished below the hoped/expected 95! Similarly if you knee jerk reassessed Selwood last season after his little mid season rough patch of 539 points in 6 games between rounds 5 and 10. It would have been so easy to knee jerk react and say "He's off this season, he will be lucky if reaches 110 for the season!". The bottom line from all this is, trust your judgmemnt, particularly on players that have proven scoring patterns, like Birchall and Selwood had. Don't rush to reassess them on the back of a hot or cold streak. Nearly all players have at least one of each each season! It was factored into your preseason assessment, so just because one of them hits early in the season is no reason to change opinion! In some cases you can see reason for it. Joe Blow's scoring has really dropped off this year, but I've noticed he has had little or no midfield time this season, and seems to be playing off the half back flank". when scenarios like this occur, you need to ask "why is it happening?" and "will it continue?". The answer to those questions are the most important thing, before you rush to a large reassessment.
This also brings me back to another discussion point that I had with a friend of mine:
Player A and Player B both had a pre-season PIT score of 100 pts/game and will both play 22 games
Player A scores 150 twice, while player B scores 50 twice. What is there PIT average for the remainder of the season. It looks like according to your theory Player A will have a PIT average of 95, while player B has one of 105. This looks wrong to me. I would say that both have a PIT average of 100 with a slight positive bias on player A and a slight negative bias on player B. It is a little bit like tossing a coin. Even so the coin came up heads 10 times in a row, the odds for it coming up tail on the 11th throw is still 50%.
Could you please comment on this.
It looks wrong, it feels wrong, and it's the hardest urge to fight, but unless you can find some underlying factor, back them both to head back towards that 100 average. I will put a rider on this notion. One season does not a scoring pattern make. Look how quickly we all wanted to label JPK a 115 player. Also, the lower quality the player, the more variance not only in their weekly score, but also their season score. The notion is better applied to players who have established scoring patterns, not 40 game players coming into their 3rd or 4th season (hence why I might have been a bit harsh with my "Zorko this year, anyone?" call, as he hasn't played enough to establish his history yet). If both those players have established scoring patterns, and there is no obvious reason for the variance, back your judgement, and say a correction will come. You can't use coin toss logic here. What you need to do is not isolate games, but look at the season as a whole. You know they will both experience ups and downs during the season, and when one has occured, history shows unless the players role, fitness or capabilities have changed, the opposite will surely follow at some stage. The good SC Coaches are counter-intuitive. Dimma's Cloke trade was counter-intuitive. On face value, and on a game by game assessment, it was "wrong". On a season basis it was spot on. Was it luck that Cloke came out and spudded up a 34 that same game? Probably. But the correction had to come, and corrections are a season phenomenom, not a game thing.
Look at your coin toss scenario, and let's assume all factors are fair, and it is a genuine 50/50 whether it comes up heads or tails. You are spot on that the 11th toss is still 50/50 for the next throw to be heads or tails, BUT ..... don't isolate your thinking to one bet, and one throw! If the coin and it's tosser are genuinely fair, there will be a regression to the mean. On an infinite scale the quantities of heads will even out with the number of tails. That would mean if you were going to toss that coin ad infinitum you'd be ever so slightly better off to back tails from now on, as there will come a point where the counts for each result regress to the mean. If it doesn't, there is some outside influence affecting the results. The more tosses made, the more likely there will be a tails hot streak in there somewhere!