[Flying Dutchman]

Let's say I enter the Hilton NFL contest. That's where you have to pick 5 teams each week during the regular season along with approx 250 other contestants.

How good do I have to be to win in 50% of the time I play?

Do I have a realistic chance of winning if I can only win 60% (51 of 85) of my picks?

Anyone have any idea how to calculate this?

[paw]

Good luck bro.............There are some HUGE Hitters that enter that contest. To beat Fezzik alone would take a small miracle.

[Dark Horse]

If you hit 60% you will probably end in the money, but you will not win the contest.

LAS VEGAS -- there's a major change in the Las Vegas Hilton SuperContest this year.
In the past, the 5 NFL selections that each contestant is required to make have been due at 1 p.m. Friday (unless you use a Thursday night game, in which case all five plays were due by the scheduled kickoff time of the first game used), but starting this year the Hilton will have a submission deadline of 11 a.m. Saturday (with the same caveat in place for Thursday or Friday games).

SuperContestants have asked for a later deadline for years in order to have more time to decide on their plays, especially to get the later injury reports. The Hilton has been reluctant to do this partly because a big reason for hosting the contest is to have people come in to pick up the sheets with the selections and the earlier the better as far as getting them to make wagers while they're on location. In a conversation I had with Hilton director of race & sports book operations Jay Kornegay a few weeks ago, he said that while the later deadline might hurt business in some ways, he's hoping that having people come to the book on Saturday afternoons with a lot of college football action going on would be a good thing and maybe result on more handle not only in the NFL but on college games as well.

The entry fee remains $1,500 with a maximum of two entries per person.

Registrations will start being taken on July 31 and continue until 11 a.m. Saturday, Sept. 12 (of course, those using the opening Thursday night game will have to register before that). The early-bird deadline, which allows players to get in a bonus $10,000 mini-contest over the last three weeks of the regular season, will be 4 p.m. on Monday, Sept. 7.

This should give you an idea:
http://www.therxforum.com/showthread...Hilton+contest

[Flying Dutchman]

OK, my analysis shows that I have a 50% chance of winning if I hit 56/85 picks (66%)

Approx. 7% if I can hit 51/85 (60%)

If I can finish with 49/85 (57.6%), I have a 50% chance of placing in the top 20.

Anyone have any better numbs than these?

[tomcowley]

You have to run above expectation to win. Of course, the better you are to begin with, the less above expectation you need to run, and the more likely you are to do it. Play with binomdist in excel and see the odds of various projected winrates beating other projected winrates over an 85-game sample. You also can't assume that everybody you're up against is a coinflipper or has independent picks (sometimes injury news will make a game an autoplay at the contest line, among other correlations).

[Flying Dutchman]

Thanks Dark Horse, a later start would help no only me but also those other slugs (Fezzik included) that are playing. BTW, any idea who's plays Fezzik is using? Can't be his own, he's not that good.

[Dark Horse]

Fezzik is good.

[Flying Dutchman]

Originally Posted by Dark Horse

Fezzik is good.

Fezzik uses other folk's plays, he is not an originator. He'll tell you that himself.

[BigCap]

If you want to win, according to my experience in these types of events and some simple modeling you need to pick about 63-65% for 200-250 contestants. Of course that figure will need to go up if you are picking against stale numbers, perhaps 2-3% points. 59-60% should get you in the top 10% of the pool. (assuming stale numbers, i.e. openers on Saturday)

[Dark Horse]

Originally Posted by Flying Dutchman

Fezzik uses other folk's plays, he is not an originator. He'll tell you that himself.

In general, or for this contest? Deciding on selections by others could be just as challenging as capping games.

I would look at it from a five or ten year perspective and consistency. Success rates can differ sharply from season to season, but if you can end in the money once or twice every five years, that should be worth it.

I guess I have to run some numbers now. lol

[Flying Dutchman]

Originally Posted by BigCap

If you want to win, according to my experience in these types of events and some simple modeling you need to pick about 63-65% for 200-250 contestants. Of course that figure will need to go up if you are picking against stale numbers, perhaps 2-3% points. 59-60% should get you in the top 10% of the pool. (assuming stale numbers, i.e. openers on Saturday)

What? The numbers are always stale. They're posted on Tuesday and don't change.

My numbers are based on 250 players, basically flipping fair coins for each pick, monte carloed over a bunch of runs. I also have a small group of sharps picking up to 55% (5% of the players).

But if anybody has a history of the Hilton, we can glean something from that. They started in the late 80s, I think.

[BigCap]

Originally Posted by Flying Dutchman

What? The numbers are always stale. They're posted on Tuesday and don't change.

Then I would say you need to pick 66-68% to have a better then even chance of winning the contest, and around 59-60% of getting in the top 20.

Originally Posted by Flying Dutchman

My numbers are based on 250 players, basically flipping fair coins for each pick, monte carloed over a bunch of runs. I also have a small group of sharps picking up to 55% (5% of the players).

Nobody will be "coin-flipping", if the numbers are stale. Everybody will have a 2-3% edge on that alone.

Originally Posted by Flying Dutchman

But if anybody has a history of the Hilton, we can glean something from that. They started in the late 80s, I think.

I have no experience with the Hilton contest, so sorry I can't help you there. But $1,500 is not a big entry fee and I would not expect their picks to deviate much from random expectation, with a peak of about 54-55% success rate for the top players.

[MonkeyF0cker]

Generally, you've needed to be up around the 70% mark to win it.

[Dark Horse]

Originally Posted by BigCap

I would not expect their picks to deviate much from random expectation, with a peak of about 54-55% success rate for the top players.

55% won't get you anything in the Hilton. A large number of SBR players pick that over a season in the BTP contest.

Not only do you need to be good, but you also need to get lucky to win the Hilton. The difference between 60% and 67% is huge, but the next season it could be reversed for the same players.

A big challenge is a reliable proxy. They're ready to offer their services, but how do you know they won't oversleep on Saturday morning? Would be great if SBR could set up a proxy system. I would commit to play for ten out of twelve years.

[Flying Dutchman]

Using "stale numbers" to give the average picker a 53% advantage the average winner hits 58/85 = 68%.

to be assured of winning 90% of the time, you need to hit 61/85 = 72%

Finish in the top 20 50% of the time is 52/85 = 61%

rough numbers.

[MonkeyF0cker]

If you need a proxy, I can probably get you a reliable one (or two), dh. I live in Vegas...

[Flying Dutchman]

Originally Posted by MonkeyF0cker

If you need a proxy, I can probably get you a reliable one (or two), dh. I live in Vegas...

Monkey, any idea of the rates?

I recall Andy Iskoe would run Hilton numbs for you a "nominal" fee.

But I forget what the fee was.

[MonkeyF0cker]

Usually around $250 or so with a 1% cut on winnings.

[BigCap]

Originally Posted by Dark Horse

55% won't get you anything in the Hilton. A large number of SBR players pick that over a season in the BTP contest.

250 contestants with an "expected" 54-55% success rate means you would need to pick at least 70%+ to win it. So you are confusing what the best person picks vs. the expected rate for everybody. If you read my post I indicated you need to pick AT LEAST 66-68% to have an even money chance (given stated constraints), which apparently coincides with history of the event.

[Wrecktangle]

Originally Posted by BigCap

250 contestants with an "expected" 54-55% success rate means you would need to pick at least 70%+ to win it. So you are confusing what the best person picks vs. the expected rate for everybody. If you read my post I indicated you need to pick AT LEAST 66-68% to have an even money chance (given stated constraints), which apparently coincides with history of the event.

BigCap, if you've got history, let's hear it.

Dutchie has shown his numbers, show yours.

[Pancho sanza]

Originally Posted by BigCap

250 contestants with an "expected" 54-55% success rate means you would need to pick at least 70%+ to win it. So you are confusing what the best person picks vs. the expected rate for everybody. If you read my post I indicated you need to pick AT LEAST 66-68% to have an even money chance (given stated constraints), which apparently coincides with history of the event.

I cant seem to recall anyone reaching 70 % in that contest, granted I only followed it a few years.

Most of the winners were in the mid 60's ATS % wise.

[Wrecktangle]

Originally Posted by Pancho sanza

I cant seem to recall anyone reaching 70 % in that contest, granted I only followed it a few years.

Most of the winners were in the mid 60's ATS % wise.

This is my recollection too, I don't remember anyone hitting 70% even once. Early on there was a prize if you hit 65%, but that was picking 7 teams per week.

[Pokerjoe]

I disagree with the presumption that stale numbers are worth 3% here.

First of all, in the NFL, numbers don't often jump around that much.

Second, they're as likely to jump against you as with you.

Third, they're only one variable.

I think they may occasionally provide a default pick for a player unable to otherwise handicap 5 picks in a given week. And they may provide the difference for someone on the bubble between two pick choices.

But it isn't going to be a generic 3% advantage.

More likely, if you handicap 6 picks, one of which now gets an extra point from the stale lines, one of which gets a 1/2 point, two of which are unmoved numbers, one of which is now 1/2 point against you, and one is now a full point against you, then you drop the pick where you've lost a point against current spread, and use the other five.

So in such a scenario, you'd have picked up a net of 1 point in 5 picks.

I think anyone entering such a contest needs to be very much at peace about the fact that the variance of return is going to increase exponentially, because not only does he face his own variance, he faces that of many opponents.

[Flying Dutchman]

Poker, I like your logic and I think it is much closer to the truth than the 3% adv. So call it a 1.15% advantage?

[Pokerjoe]

Something like that. I don't think real precision is possible, tho, or necessary. As you've pretty much already done, you can approximate a field of mostly 51% players with a small mix of 55% players, then see the distribution, and then cross-ref it with your own distribution to have some idea of your chances.

But here's a good example of how much luck there is in it: the one year I got to enter (have never trusted proxies) I 3-way split second with about 57%. That was about 20 years ago, btw.

[Flying Dutchman]

Originally Posted by Pokerjoe

Something like that. I don't think real precision is possible, tho, or necessary. As you've pretty much already done, you can approximate a field of mostly 51% players with a small mix of 55% players, then see the distribution, and then cross-ref it with your own distribution to have some idea of your chances.

But here's a good example of how much luck there is in it: the one year I got to enter (have never trusted proxies) I 3-way split second with about 57%. That was about 20 years ago, btw.

Poker, which year? I was in the original '87 & did '88 & '89.

Was tied for the lead in '87 thru the strike, model when sideways when everyone got back together. If I remember right, only about 100 players was in the first after the old Castaways tourney folded.

[Bogart45]

I'm sure I read somewhere that nobody has ever picked 70% in this contest. You get a 10k bonus if you get higher than 67% and that is almost never won. 65% should get you at least top 3 and 61% get you in the top 20. This is based on actual results which you can search for on the web.

[cankid]

friend of mine that won it a few yrs ago went 58-25-2 and collected also on the $10,000, unreal yr he had

[cankid]

If you or anyone you know needs a proxy for the Vegas Contests contact me. In the past when I lived north I had a hard time finding anyone to proxy.....I network with some of the best sports guys in the country including 2 of the last 5 hilton winners......hope to see some of you in a few days