“Small or Far Away” Father Dougal on Cows
Since this is the round for correctional trades, it is the round to write about correctional trades. By correctional trade I mean one that is not forced by serious injury. Fixing a mistake, or fixing what looks like a mistake due to under-performance.
But first, my once a year bit about the origin of “Small or far away” and the staggering coincidence of Father Dougal writing about cows: Small or Far Away
(That green highlighted bit is a link, in case it wasn’t obvious)
So, corrective trades…..
Very short version: Bad (Tree pretty.)
Short version: Very often not worth it, be very careful, especially with rookies.
Full Sermon: To start, there is this thing called “regression to the mean.” I’m probably going to butcher this but basically, when you have a result far from the average, future results are likely to be closer to the average. If Buddy averages three goals a match and kicks ten in a match, odds are he will kick fewer than ten next match. If you prefer something less skills based, if you roll two dice and get 11, you are likely to get a result closer to 7 on your next and future rolls.
Last season for the R1 Cow Talk I looked at the 2016 cows and how their scores were after the early round and the rest of the season. Link below if you have not read or don’t remember it. Oh, and I figured out how to do a normal click-on-the-line type link!
To sum that up: improving by early cow-trading looks to be difficult. Especially when you include the cost of the trade.
This season, I’d be inclined to think Kelly, Holman, Deodee, and Finlayson are worth finding a way to get in, because their averages are both freakishly high and appear to be real. Kelly being the main one.
Okay, you just talked about regression to the mean and then said, oh these guys with high two-round averages should probably be brought in. Which is it? Just trying to cover your butt?
Well, um, I think there are two things at work here. One is that these guys were expected to do well, and they are backing that up. It’s the guys who come from nowhere and put up completely unexpected scores who are a big risk. If say, one of the three Eagles forwards had put up a pair of hundreds, I’d be suggesting caution. The other thing is the good old eye test. Watch them play. If I didn’t know Tim Kelly was playing his second match, and you showed me both the Cats v Hawks and Giants v Pies matches, and then asked me which Kelly was the rookie, I’d have said “neither, stop messing with me!” If I had to guess I would probably have picked *the wrong one*.
Yeah, okay. They all pass the eye test, I’ll give you that.
There’s another reason, which is that the ones who are not Holman are looking like they might actually be keepers.
Yeah, I almost put something about “won’t be keepers” in the rules of cow-quisition but couldn’t come up with one I liked. Good thing as it turned out.
BURN the heretic!
Please do not burn the heretic. Also, let’s be real. Any mid who you start with and averages 105+ is at worst a low priority upgrade. And defender you start with and averages 90+ is at worst a low priority upgrade. We’ll see how well their averages hold up before we need to decide if they should be harvested. Once you have someone on your team, their price only matters if you are going to sell them. There is no actual requirement to sell a cow. We just call them “steers” if they do that well. Would you trade out a top 8 mid or a top 6 defender?
Of course not. At least not while sober…You don’t think any of them will really do that well do you?!
Let’s say I am not sure all of them won’t, and there is no reason to try and predict. Again, we’ll find out.
So if they are even candidates for keepers then they are good enough to bring in?
Right. Remember last week and how I talked about the value of fast points?
Well, go read last weeks Cow Talk. It’s about the value of early points.
Well, if early points are valuable doesn’t that mean corrective trades are good, fixing a problem fast?
Yes, if what you are doing is actually fixing a problem. Say we are rolling three six sided dice, aka 3d6…
Yes, we all know that. Anyways, the average of 3d6 is 10.5. But if we roll 3d6 say six times, we might get – 15, 17, 12, 7, 9, 10 – which averages 11.6. But if you only look at weeks, I mean rolls, one and two, they average 16! What great dice! Clearly you should spend a trade to bring then into your team, replacing the 4d6 that rolled 11, 10 for a mere 10.5 average!
I think you are mixing your metaphors….
Quiet, I’m on a roll. Anyways, the 4d6 then go on to roll 14, 16, 13, 20 for a average of 14 over 6 rounds. But thanks to just two rounds, I mean rolls, we traded the set of dice which should average 14 for one which should average 11.5. Fewer points and a trade gone.
AFL players are not dice
Doesn’t matter, same idea
DOES SO matter. For one thing we know what the average is for dice, so since we know for sure that 4d6 will over time average 14 and 3d6 over time will average 11.5, we know that is a silly move. So nobody would do that.
Fine. Well, we do have some knowledge of all AFL players. Even rookies like Tim Kelly. To further mangle my example, by rolling the equivalent of a 22 we know he isn’t just 3d6 because 3 x 6 = 18 and so they can’t roll a 22. Must be at least 4d6. And odds are that it is more since rolling a 22 twice is improbable for only 4d6, so he is likely to be more dice then four.
You are blithering and overstimulated and should try again more clearly. Just because you know what you mean doesn’t mean the rest of us do!
Yeah, I think I went too fast. I’ll have a beer and take some deep breaths, not at the same time.
Now let’s try again.
There are random effects that go into a player’s performance, just like there are random effects when rolling dice.
Over the short term those random effects can make a player look better or worse than he really is.
Over the long term, random effects tend to even out.
So, when a player does extra well or extra poorly, and there is not a good explanation for that, it could very well be just luck, and if so, he will move back to performing at his real level of ability.
So, if we roll three dice a lot, they will average more than when we roll four dice the same number of times. But if we look only at the first two throws of the dice, sometimes the three dice will do a lot better than the four dice due to random chance. In the same way, sometimes an AFL player will have a Supercoach score well above or below their real level of ability over the first two weeks. When that happens, we need to think hard and determine if their performance is because of luck or other factors. Do they have a niggling injury? Is the coach using them in a different role? Are they old and perhaps dropping off because their body isn’t what it used to be? Did they just play two teams that let up big scores, or let up big scores to players in their role? Has their time on ground gone up or down?
Ok, that makes sense. Thus the rule against trading your premiums, like Fyfe after his bad score week one.
Exactly. There is risk both outgoing and incoming. If you get one wrong, you have spent a trade unwisely. If you get both wrong, you have spent a trade to go backwards.
But if you get both right….?
Then you have made a good trade. The belief that we can get both right is tempting. We know we can. And we really can, sometimes. I think about these every year.
And you made *two* last year, Mr “do as I say and not as I do.”
Yeah. Time for some Deja Vu. Last season I swapped Cam McCarthy for Will Hoskin-Elliot. I also broke the rule about sideways trading premiums, and do not yell “heretic” again. Got rid of Hamsterberry and brought in The Bont.
How did they work out?
Well, McCarthy was a total spud and stayed one all year. Worse, he was an expensive spud. He went 36, 28, 53, 100, 72, 78, 73, 66, 61 to start the season. And that 100 and the 70s meant he would have ended up making 100k had I sat him on the bench and not traded him.
WHE went 88, 120, 71, 70, 51 ,125, 65, 75, 71 to start the season. I ended up making about $160k from him. Meaning my return on that trade was about 60k and 25 points.
Yup. Here is what I said in R2’s Cow talk last year: “If McCarthy to WHE isn’t a good correctional trade, then there is no such thing as a good correctional cow trade. And since there is such a thing as a good correctional cow trade, McCarthy to WHE must be one.”
You were wrong
Yeah. I mean look at them after two rounds – an average of 32 and an average of 104! How obvious is that? How many of us would do that? And yet, not worth it.
So, “since McCarthy to WHE wasn’t a good correctional trade, then there is no such thing as a good correctional cow trade?”
Well, let’s say I won’t be making any more, ever. Probably. If I do I will probably suffer for it. So no. Notice me not even *thinking* about trading Naughton. Maybe next year I’ll look at all the possible Naughton trades and see how they did.
But you said to bring in Doedee and Finlayson….?
I have them, which makes not thinking about trading Naughton easier. If someone does not have them, I would have no issue with bringing one of them in for Naughton. Or Kelly if that can be worked out. I’d do that for sure.
Ok, how did Hanners to Bontempelli go?
Hamsters averaged 61.5 over rounds 1 and 2 then 101.1 over the rest of the season. The Bont averaged 130 over round 1 and 2 and then 102.6 over the rest of the season. So, ick. As it happened, Hanners missed a week and the Bont did not, so I ended up 131 points ahead, which is about breaking even. But, this was a case of getting something half right. Hannebery *was* hurt and his average dropped by 10 points from the previous season. I was right that he could be improved on enough to make a trade worthwhile. What I stuffed up was who to bring in for him.
Go on, try to claim not being wrong
M Crouch, Dusty, Titch, Oliver, Murphy, Sloane.
Had I brought in any of them it would have been a win. I waffled between Bont and Dusty. Had I gotten Dusty, I would have been 427 points ahead.
Yeah, what a difference *that* would have made at the end of the year. So if you do trade out a premium, for sure only trade them for someone you are certain will be a top scorer. I blew it by trying to save a little cash on The Bont vs Dusty, and cost myself 296 points.
So you’re saying the rule about not sideways trading premiums is a bad rule?
Nope. It is a good rule. Hanners was hurt.
We didn’t know that at the time.
Eye test. I posted about it in various threads. He looked wrong. I had owned him and seen him the previous season, and he just did not look right. And I trusted my gut, well, eye, along with the scores, and it turns out he really wasn’t right. Still isn’t right as of today actually.
So this season….?
I am not getting sucked into that! The only real question I know of is Hibberd and with him it is role and not injury. And I have no freaking idea how real or permanent the role change is. Whatever I do, I already know something that will be in Cow Talk R2 next season.
What are you going to do?
No idea! If I trade him out it will be for someone I am really sure about. Oliver if it can be worked, or Simpson if I go to another defender. Could being in a forward if there was one who wasn’t an R14 bye I wanted. I’m really torn on what to do….
Yeah, well, so are most people who own Hibberd. Start a club.
I’ll probably have to do the “which way do I want to fail thing.” Anyways, enough about my team. I wouldn’t have brought it up if those examples weren’t so useful and so many other people were not in the same boat and I hadn’t just written about that boat.
So it’s a big boat?
A very big boat?
I guess so.
A “Titanic” boat?
Expletive off. Anyways, so ends my yearly round two sermon about the dangers of “corrective” trades.
You forgot “the thing”
Oh right! This is not related to the corrective trades stuff, I just thought it was interesting. Remember all the blokes I listed as better options than The Bont? All of them as well as The Bont, averaged higher in rounds 1-2 than they did the rest of the year – other them Tom Mitchell. His average went UP after round two by 2.1 points. I don’t expect him to have an average this season in rounds 3 to 23 of 159.6…..but if you don’t own him…….
Leave a comment / Scroll to bottom