Small or Far Away – Father Dougal on Cows
First, thank you to Lesley and Weendog who found examples for me. I sadly didn’t see Weendog’s in time to save the data I need to check it, (my bad, I’m sorry,) but I did see Lesley’s in time and it tuns out the data I needed was there to start with.
“Darcy Parish was projected to score 83 and, you guessed it, scored 83.
He was projected to appreciate by 48.8k (which would leave him at 309.2k), I figure that if these projected figures are marginally false come price changes, it could (at the very least) prove your rounding error theory.”
So, what really happened with Darcy Parish’s price? It went up $48,700 to $309,100! Players hits their projected score but was off $100 from the projected price. Proof the SC gold numbers are not 100% reliable. Which we “knew” but now know for sure this season. Let’s look at some other numbers and see how that fits in.
The “BE Proj $ Change” column is what the price change at the projected score should be using the multiplier that makes the SC BE work right. To get the “BE Proj $ Change Real score” we plug in the real R4 score in place of the projected score.
We can see that even using the SC numbers within a week there are inaccuracies. We have a breakeven from them. We also have a projected score and projected price change at the score. We can take the BE and plug it into the formula to get a multiplier, but when we plug the projected score we do NOT get the projected price. We should, but for some reason we don’t. And the errors are sometimes more than $100. Both Libba and Schache are off by over $200 and in difference directions. This goes right along with what we saw from Parrish. The SC provided numbers are close but not actually accurate.
Below are the same players with their prices so we can see how they don’t match up. “R4 Price from R3 BE” is taking the multiplier that makes the BE from SC work (Projected #) and plugging in the real score in place of the BE score. “R4 Base Price no #” is the base formula with no multiplier. We divide the real price by that to get the real #. (With # meaning multiplier.) e
The nice thing is that the real #s are all very close. They are also not that far from the Projected numbers. They are not the same, which at this point we know they won’t be.
So, any conclusions?
The numbers provided by Supercoach gold are not completely accurate. At this point in the season they are a little bit high on average, but only by a few hundred dollars and they are usually within $200 in either direction. In practice, there is no reason not to use their numbers for planning purposes. I care that their numbers are kinda weird because I am trying to use them to make my own projections and do analysis, but for planning trades and such they are good enough. I’d love to have been able to say “Ah ha, their numbers are all messed up, read my articles to get the real stuff,” but, it turns out the weirdness is usually too small to sweat.
Oh, how to know when to sell cows! D’oh!
The conventional wisdom is that you should sell a cow when their BE is about the same as their average, or when their BE is about the same as their projected score. Neither of those is wrong, but I think they are a bit too simple and we should take more things into account. Ask yourself the below questions:
First, is there a really high or low score throwing off his price? If there is one really high score then selling before it goes out of the cycle makes a lot of sense. If there is a low one, the player may well shoot up after it leaves his cycle, and holding on longer can be in order.
Is there someone to downgrade to? Huge question. Early in the season you probably don’t want downgrade into a player who won’t be a cow too. Later on you will end up without enough trades to cash in any new players, and you may want a very cheap loophole guy rather than someone who will make money you can’t spend. Of course making money means playing and being a good backup, and those are good. You may even want to downgrade a little early if the player you are bringing in will make more than the player you are selling. Getting the last $10,000 from a cow is probably not worth missing a player with a BE of -75, on the bubble, and named for the next round.
Do you need the money for something? The goal of money is to buy players, and so having a little less income from a cow sooner so you can get a player you want sooner, can be worth it. If the player you want goes up by $30,000 it does not matter if you get an extra $20,000 from selling a cow, especially if you then can’t afford the guy you want to trade in. And the longer you wait the fewer points you get.
Has he lost his job? If so his value will not change, either way. Maybe a good loophole while he waits for another chance. Or if he is injured long term, you can cash him in when you have a downgrade target and do not need the loophole.
Are you going to make a decent amount of profit? If he is at $170,000 for example, he may just need to sit the rest of the season. No player is cheap enough to make a trade worth using on someone in that price range. (Unless you really need $65,000, which could happen.) Hewett is in that bad place. Not worth selling, not close at this point, but not playing. Loophole until he plays more.
I don’t know really the money cost of a trade, but to me last season trades ended up worth 300 or so points. The sooner you run out, the more a missing trade costs you. Last round can cost 100+ points not replacing a premo. Second to last round, 200+, third to last, 300+, etc. Trades are really valuable, so, don’t use them to make a small amount of cash unless that small amount means big points.
I’m sure there are more things to think about out there, so please share your thoughts in the comments. Cow selling is not a precise data driven decision, there are lots of factors that go into each decision, so please share any thoughts or stories you have.
Leave a comment / Scroll to bottom