Topic: 2014 USD/mBTC Price Prediction Contest - page 14. (Read 21790 times)

If no new money, and 4 or less of the current entries remain until the last round (remaining does not cost money, but you should not make blunders), the winner takes it all:

Winner: 1,400mBTC for a 28x payback!

If 5-8 entries remain in play, the pot will be distributed as follows:

Winner: 1,060mBTC for a 21.2x payback!
2nd place: 340mBTC for a 6.8x payback

Prediction sent.

If I was now doing the exercise of filling 8*80 fields, I would do it the following way:

1. Forecast May first as it is only 24 days away and historically a violent turnaround (going to 1.00+) is less probable. The total destruction of price is also less probable according to the trendline model and the clear tendency of the price to snap back. So we would likely stay in 0.25-1.00 range centered on the current price.

2. Forecast December second. The center of the distribution would be where you think the trendline is and where we are relative to the trend. It is quite possible that we are in the dumpster (3) after having had a moonshot (8 ) in between. All the prices from 0.5 to 50 should receive some coverage. (If this competition was started in November -12, who would have guessed the yearly advance of 100x?).

3. The months in between are just interpolation, with perhaps a slight tendency for moonshot in the months when I think it is most probable.

* *

We have now 8 entries competing and Round 1 entry is CLOSED. I am waiting for all to submit their predictions today.

At the close of May, as round 2 starts, some more players may enter in since they know about the system much more and see the good odds

filling in all the fields is quite a task

But I'm almost done... looking forward to the results

I send in an initial prediction, I might review it and send in a new one (to override this one) if I have time.

If somebody is hell-bent on winning the contest by taking excess risk, he may get eliminated by that rule.

But anyone who is not, should have a decent chance of ending up among the top 25%, netting at least double his money back.

It is 14:42 hours to go until all entrance fees must have been paid. After that, there is an additional 24 hours for submitting the prediction.

I am also not a pro I guess. I am too bullish biased
But hey, I can work with regression (excel / SPSS) so I will try to have "neutral guesses"

And if I am eliminated by the -20 rule, I will enter again

I'm not a hardcore pro, either. I've only ever done softcore stuff. Oh, wait, you're talking about trading. Never mind.

You might be right but the payout structure of the competition is top heavy therefore I expect participants to take (calculated) risks in their prediction. It's better to optimize your chances of coming in first than to optimize your actual score. (This makes the optimal tactic comparable to the end game of a poker tournament).

Anyway, I am participating so at least one of the participants so at least one is not a hardcore pro (I never traded anything, value investing and sometimes a bit of low risk arbitrage (outside of Bitcoin)).

Until now, all participants are hardcore pros, so if an entry fails by -20, it is likely a calculated move and hedged by another entry

Getting such a bad score would require that something unexpected happens, and a large share of the others somehow know it and you don't. The first one alone is not sufficient.

Note that in the first column of the amber cells, there is the historical probability distribution for the 30-day change centered in 447.

I have the feeling more than half the field won't make it to the finish line.

Well it will be incredibly difficult for everyone to get to -20 if the average score per round is zero anyway... 

See the final ruling above, in most cases that should leave a few entries standing, who will share the pot.

Sounds very fair

To make sure that all of the prizes go to participants, this is the final ruling:

If all remaining entrants are simultaneously eliminated (disqualified) in round 8, all the disqualified entrants from round 8 (but not from previous rounds) share the prize pot normally, and the competition ends.

If all remaining entrants are simultaneously eliminated (disqualified) in round 7, and there are no new entrants for round 8, all the disqualified entrants from round 7 (but not from previous rounds) share the prize pot normally, but the competition ends, so there is effectively no round 8.

If all remaining entrants are simultaneously eliminated (disqualified) in round 6, and there are no new entrants for round 7, the competition continues with round 7 being an empty round, but there is a possibility to submit a new entry for round 8. These entries are in competition for the prize pot. If nobody chooses to enter despite the good odds, the round 6 disqualified results are used as a basis for distribution.

The same for rounds 3-5.

In rounds 1-2 it is not possible for all the players to simultaneously suffer -20 points.

As a participant I would prefer that in the even of a round where everyone is eliminated the final score _at that time_ is used to determine the winner. Alternatively, the scores at the beginning of the round could be used to divide the prize pot.

For the record: I will not object to your proposal above

Wouldn't it make more sense to just not eliminate anyone if there would be nobody remaining, and award the prize to whoever failed the least miserably?

I was starting to write that it is impossible for all players to be eliminated simultaneously, but actually proving it is rather difficult. I'll try anyway:

- Let's assume all players are infinitesimally close to being eliminated after round 7
- If there is zero players playing a certain month, it gives zero points to all non-players.
- If there is only one player, he likewise gets a zero regardless of the prediction, because the average 100% consists of him.
- For two players, one gets points equal to -(the other). The total is zero (still impossible to eliminate both at once). But...

- For three players, it is possible to get a negative sum for points awarded for that month. Eg. if one player votes 1%, and the two others predict 10%. The result is -3.6416 points for the worst one, and 1.1544 for the other two. The sum is -1.3327 points.
- Therefore, if three players remain in the contest and they have several predictions that are evaluated in the Round 8, and each of them has failed miserably in one of the predictions, both the individual month sum totals and all the players' totals for that round CAN indeed go negative, resulting in elimination of all of the remaining players.

I don't see any other way to keep it equal opportunity for all, except to award all prize pot to Malla Satoshi Suite construction effort if this happens.

I've manged to get Bitcoin-Qt running in a chroot jail off a Live-CD (my initial plan was to rollback my drivers that way, but apparently apt-get and chroot don't get along for some reason I'm still trying to discover).

BTW small question concerning the price allocation:

What if no-one reaches the end of round 8 due to being eliminated because of a -20 score? Will the prize pot go the last man standing? Will the money be divided equally among all? (Will Risto buy a very nice cigar?)

Sounds like you have a fun Sunday ahead of you

Anyway, good luck, I hope you resolve it quickly.