So I'm thinking of doing a spin-off of this for the summer contest, andI'm wondering if the idea for it seems sound. The scoring system wouldbe the same as the one currently being used here, but the twist wouldbe that predictions for all 63 matches would have to be submittedbefore the start of the tournament...think of it as Oracle Challenge: Extreme!!!!!!

Thetrick would be scoring any matches after the first round where peoplemight be missing one (or both) of the combatants. I considered allowingpeople to simply make new predictions in exchange for taking a bit of ahit pointwise, but it would take away from the fun of it all if peoplewere allowed to base any of their contest predictions on results fromthat contest. The best method I've come up with so far would be to useextrapolated predictions for those types of matches. I can't quiteexplain what I mean in words, so I'll use a couple of examples:

********************************

Example 1:

Contestant SuperNoob makes the following predictions for a four-pack in his bracket:

Mega Man over Mr. Game & Watch, 84.90% - 15.10%
Gordon Freeman over Yuna, 51.00% - 49.00%
------
Mega Man over Gordon Freeman, 63.00% - 37.00%

Inthis example, let's assume that Yuna actually beat Gordon Freeman52.00% - 48.00%, and that Mega Man won his opening round match. SinceSuperNoob believed that Yuna's strength was .98 that of Gordon Freeman,his prediction for Yuna's second round percentage would be computed as.98 * 37.00, or 36.26.

********************************

Example 2:

Contestant StatsBoi makes the following predictions for a four-pack in his bracket:

CATS over Kuja, 60.00% - 40.00%
Andross over Ghaleon, 53.00% - 47.00%
-------
CATS over Andross, 55.00% - 45.00%

Inthis example, let's assume that Andross won his opening round match,but that Kuja beat CATS. Since StatsBoi believed that Kuja's strengthwas .80 that of CATS, and that Andross's strength was .90 that of CATS,his prediction for the second round match becomes Andross over Kuja,55.56% - 44.44%.

******************************

Example 3:

Contestant BlIzZaRd RuLz makes the following predictions for a four-pack in her bracket:

Diablo over Shadow the Hedgehog, 52.50% - 47.50%
Sarah Kerrigan over Lloyd Irving, 55.00% - 45.00%
-----
Diablo over Sarah Kerrigan, 60.00% - 40.00%

Inthis example, let's assume that Shadow the Hedgehog beat Diablo, andthat Lloyd Irving beat Sarah Kerrigan. BlIzZaRd RuLz believed that thecharacters rated as follows:

Diablo - 50.00%
Shadow the Hedgehog - 47.50%
Sarah Kerrigan - 40.00%
Lloyd Irving - 36.00%

Therefore,the prediction for her second round match would be extrapolated asShadow the Hedgehog over Lloyd Irving, 62.11% - 37.89%.

*******************************

Theone problem I can see with this method would be accounting for SFF. Foran extreme example, let's say a four-pack read something like:

Link over Duke Nukem, 83.00% - 17.00%
Ganondorf over Auron, 52.50% - 47.50%
-----
Link over Ganondorf, 85.00% - 15.00%

Thatperson would be in serious trouble if Auron beat Ganondorf, becausetheir second-round projection would become Link over Auron, 85.75% -14.25%.

I suppose that I could let people submit "what-if"scenarios should anything like this show up on the bracket, as thereshouldn't be too many instances where something like this could occur.

******************************

Anyway...for those who read through the muck:

1) Is it an okay idea?
2) What improvements would you suggest?
3) Is it something you'd consider participating in?
---
ALL YOUR VOTES ARE BELONG TO CATS