Cow Talk – R2–>R3

Written by Father Dougal on March 31 2021

Hi Everybody!


So at this time every year I talk about not overreacting to just two weeks of data.  How most players who are averaging over 120 are going to average a lot less by the end of the season. How a lot of players are going to end up averaging a lot more by the end of the season then they are now.  The whole small data samples and regression to the mean stuff.

How about if I am new to this? Don’t new people deserve to hear about those things?

Yeah, well, fair. I will sum up. 

Small data samples are when you have so few pieces of data that you can’t count on what you have being an accurate representation of something. 

Say you roll two dice 500 times and then you roll them 500 more. Odds are high that they will average about 7, which is what two dice should average. The more you roll, the more likely they will end up averaging 7.  But let’s say you roll those dice twice and get a 12 and a 10.  That averages 11! Those must be the dice to get if you want to roll high, right? Well, no, you just happened to roll a 12 and a 10, which are both possible. So the dice, being true, will average 7 or very close to given more changes, meaning a bigger sample size. 

Are those dice named “Tex” and “Walker” by any chance? 

Well, I guess they are now if you want. I don’t name my six siders.

You name your twenty-siders though.

Well of course I name my twenty-siders! I just don’t name my six siders, I’m not a psycho!

Anyways, small sample size means what looks like a good or bad start might really be just random. If it is, they will probably go back to averaging what they should for the rest of the season. If you plan to roll two dice twenty two times, they should average 7.  If the first two rolls are 21 and 10, then the next twenty should still average 7, but the average for the whole 22 would be 7.36.

Shouldn’t the dice roll lower so they end up averaging 7? 

Nope. That is not how it works. They are expected to average 7 from here on out, but the previous results have no effect on that.  The idea that past results will affect future results is called the Gambler’s Fallacy.  

But AFL player’s are not dice!

Obvious and yet a great point.  We don’t know for certain what an AFL player “should” average.  We might have ideas but they are often wrong, and we can’t for sure tell the difference between a player who average 100 because they should average100, and one that averaged 100 but should have averaged 90. That makes it fun!

So, bit of a long sum up, sorry. 

What I really want to talk about is trying to put what has happened so far this season into a good context before making any moves. 

You forgot the making moves is bad unless it isn’t bad part.

Yes, um to sum up again. Trades are very, very valuable. You need to spend them wisely. I think of trades as oxygen, if you run out before your need for them runs out, it goes badly.  That’s why it is so important to trade wisely and only when you can really benefit. 

So, that context thing. 

Hopefully, before we selected our teams, we thought about who we expected to do well and why. Also who we did not expect to do well and why. Maybe not for every player, but for lots of them. What matters as far as trading in and out, is not as much their scores, as did anything fundamental change with that player to change our assessments.  

Isn’t someone’s scores being way off what we expected a clue?

Sure! I would at least want to look into any player I own who is way underperforming, and all the ones I don’t own who are way overperforming. For examples, Cripps. If Cripps looked fine and scored an 88 & 82, that would suck for his owners, but there also might be reasons. Extra frees against, or bad DE, or just things not going his way. That pretty much calls for holding him. No point in spending a trade on someone who will come right back and score well for the rest of the season. 

Like Grundy after round 1?

Yes, exactly! 

But, if you watched Paddy play, in both games he earned a “Hank Hill” 




He did not look right and it didn’t look like luck.  We were guessing something was wrong based on how he looked not right. It turns out he is playing hurt.  That information means he is not an automatic hold, but a possible out.  Same with Neale. There is a reason his low scores are a clue to a real change in ability, at least for a while, and that means he goes from hold to trade. 

That clue thing can go the other way too. If say, Callum Mills and Andrew Brayshaw put up two  scores well over their previous averages, that is evidence of an expected / hoped for change.  Mills did move from defense to the mids, so he is scoring more. Brayshaw could be expected to improve and do better with more Fyfe up forward, and he is. Will they both end up averaging over 120 the rest of the season?  I’d be surprised. Are they both real chances to average 110-115 from here on out? Yes, and their having done well is some evidence of that. 

But, speaking of Brayshaw, last season he started off really badly, with a 78.5 average over his first four games. Many people jumped off, including me. Then he went for 106.4 for the rest of the season. If we had picked him for a 100+ average for the cost of a 70.5 average, and backed ourselves, we would have gotten a 101.3 average guy at a huge discount.  But, had to stick with the plan in the face of a scary start. 

Jordan Ridley is an example of the other sort of fellow to watch for.  I don’t think many people expected him to break out last year, but suddenly he was. Like Oliver his second year. Anyone with the guts to grab him early did very well from it. I’ll note players who actually sustain jumps like that are usually young or even very young.  Old is not impossible but if you are going to back an old guy to make a big jump in output, have a very good reason. 

There is no rule that you can safely follow. Most of the rules are really more like  guidelines. Don’t sideways premiums – Unless you have a darn good reason.  

Oh and as this is Cow Talk, on cows.  Bringing in really good cows is often worth a sideways trade. We have three cygnets and one imp that look worth getting in at the cost of a trade.  Maybe some other guys depending on your situation.  Well, Flynn, but we all have him, right? And maybe Highmore maybe. You have to make a lot of cash from a traded-in cow to make up for the trade you used to bring the him cow in. Generally going from a bad cow to a good cow is worth it Going from a medium cow to a good cow, or a bad cow to a medium cow is not worth it. Needs to be two steps up. 

Two step? I thought people danced with Wolves, not cows?

I thought it was waltzing with Bears?

Oh! Oh! How about Hambo with Hamsters!


It’s Swedish and traditional!

Moving on,…um, don’t make small moves with cows. Big ones yes, small ones no.


Dodgy Advice:

Bringing in Tex Walker as a cow-like Object? I mean, that is really dodgy advice, but I fear this time that is actually just bad advice. Of course I may just be having flashbacks to 2015, when a still mostly clueless me brought him in after his 177 in round one, and then, um, I can’t think about it.  It was bad, very bad. I leave that up to you all

Is bringing in Tex Walker as a Cow-like Object dodgy advice

  • Yeah, pretty much dodgy all right. (39%, 94 Votes)
  • No, that was bad advice! Tex is a Bum Steer! (35%, 84 Votes)
  • No, that was good advice! Moo Tex Moo! (26%, 62 Votes)

Total Voters: 240

Loading ... Loading ...

Well, just remember “You knew he was Tex Walker when you took him!”


Remember the Cow Talk guarantee: “All predictions wrong or triple your money back!”  (Offer not valid if money is actually involved.)

I am time zonally challenged.  When Cow Talk goes live, I’m probably asleep, so replies from me may take a while.

Thanks for reading!




Leave a comment / Scroll to bottom

15 thoughts on “Cow Talk – R2–>R3”

  1. Is there a 2 step up in Dow to Impey?
    I consider Dow to be a bad-medium and Impey to be a medium-good cow. Could be a two step, could be a zero step. Not sure it’s worth a trade or not but would be my only trade so far if I did (before teams drop). Anyone else with thoughts on this?
    TU – Worth trade
    TD – Not worth trade


  2. Insightful as always FD!

    This year, I made sure to employ your strategy of picking all the good cows, and then cow-like objects before picking keepers.

    For me this included:

    DEF: Clark, Highmore, Koschitzke

    MID: Dow, Campbell, Powell, Gulden, Berry, Brockman, Scott

    RUC: Flynn (wasn’t ballsy enough to roll the dice at R2 with Meek as backup though!)

    FWD: Ziebell, Daniher, impey, Warner, Rowe

    Now, I am very happy with all my choices except I am sure I have missed a good cow in Jordon, who I originally dismissed due to JS, but Harmes’ injury and good performance has seemingly locked that away for the time being.

    However, while this may be a hindsight call, upon reflection after round 2 I feel I took 1 cow too little and obviously compensated by taking 2 seemingly bust picks in Phillips and Heppell. Phillips I saw as an underpriced keeper and Heppell was my last picked as he simply fit the price point, and barring injury I saw him as a fairly bankable 100-150k cow-like object. I plan to split them into Jordon and Ridley this week, as much due to fact it allows me to field another cow with high scoring potential (fielding Rowe over Warner this week was definitely not a pleasing decision!). My current state would have me benching Dow, Rowe and Berry – all of which seem to have decent scoring potential and with the manipulation of match ups and use of a loophole on occasion seem fine field options.

    So now to my question, do you think it is worth taking an additional seemingly dodgy cow from round 1 to the ones we consider good, with the foresight that at least one unexpected good cow will pop up early, as mathematically speaking guns and rookies is ideal to the extent that it remains practical with the picks we are given by the SuperCoach gods? Or, do you think it is just as likely those seemingly good cows will not turn out so good and may be dodgy (I am looking at you Scott!) and thus require a correction on the eve of round 3?

    Would definitely appreciate your two cents on this conundrum for next year’s starting team!


    1. Oh, great question!

      I also missed Jordon for the same reason, and plan to bring him in. I’m far from thinking Heppel and Phillips were bad choices or failed. Heppel scored well and got hurt. Phillips has had a bad scoring game R2 because he went clanger happy, but still got a lot of the ball and won’t do that every week. I have him and plan to hold him, and expect he’ll still end up fine long term. He does not have to do that well to be worth what we paid.

      Nothing wrong with bringing in Jordon and Ridley for them if you want to though, extra with Heppel hurt. I probably would not trade Phillips on his own, but as part of that combo I can see it.

      But for your question: Rule of Cow-quisition Number Six-A: “A cow and a premium are better than two mid-pricers with the same total cost.”

      So, if there is a good enough cow, then follow the rule. if not, then the rule does not apply.

      Also, if you go the no starting loophole route, you can risk a cow with uncertain JS more safely.


      1. Very true! Although I have decided that a loophole is worth it unless there is obviously too many amazing cows, especially with 24 hour team announcements again this year and the introduction of the medical sub – you could get caught out with a cow that has lost his job but is on the fringe.

        So in essence, if the same situation was to arise next year, I should back in the same plan and still pick the same cows in round 1, from which point you can always adjust after round 2 if more pop up? Because next year could pose the opposite conundrum: with seemingly good cows losing their job and new good ones popping up?


        1. You pretty much have to go with what the best option is each year and it will change every year as the players change.


        2. The biggest change is the medical sub. it has removed certainty on loopholes but for all bar the most deliberately selected ones.


  3. In previous seasons bringing in a key forward is dumb advice.

    But in 2021, bringing in a key forward who can legally waltz 10 metres closer to goal without interference or the man on the mark moving?

    Different story.


  4. Great stuff, Father! The stuff about small sample sizes (and the Gambler’s Fallacy) should be mandatory reading at this time of year, much like the Rules of Cowquisition are in the pre-season.


  5. Great write up once again Father!

    Originally traded in Tex for T.Green (via DPP) earlier this week but getting cold feet despite the Crows’ run of fixtures. Thinking it might be wiser to opt for Warner instead who is the only other cow I’m missing. What do we think?

    TU: Tex
    TD: Warner


  6. Always a good reminder, thanks. While reading this I had a “its too early for thinking” thought and may be a bridge too far (or already done?)but here goes.

    If we exclude last years data, what would be Ave score for the top 6 defenders, 8 midfielders, 2 rucks and 6 forwards over a data set of 3,5 and 7 (is 10 relevant?) years for trend analysis. I’m not a stats guy but I think having a target metric for each position would be very interesting in how we value players and trades.



Leave a Reply

Your email address will not be published. Required fields are marked *