Forum: foot
Page 3697
Subject: Football Pickoff - Ideas for 2004/05

  Posted by: Guru - [330592710] Mon, Feb 02, 2004, 11:41
Before we forget about it, feel free to leave your thoughts about possible improvements for Football Pickoff for next year.

I know there are some ideas buried in the season discussion thread, but I thought a separate thread was a good idea.

Among the ideas that are worthy of consideration:
1. Changing the bonus formula from a flat 50 points per game (for 10 or more correct picks) to a graduated scale - perhaps something like 10-20-30-40-50-60-70. Or maybe even 10-30-50-75-100-125-150.

2. Changing the tiebreaker criteria. Eliminate the "most correct winners picked" tiebreaker. (See post 37 in the thread linked above.)

3. Include a prize for the top result during the playoffs, just to encourage greater participation during the playoffs.

Are there other ideas that could improve the game? These could be related to rules, prizes, or even site features.


 
1Sludge
ID: 150192911
Mon, Feb 02, 2004, 12:23 		With regards to #1, I kinda prefer a 20-30-40-50-60-70-80, as this keeps the number of bonus points possible for a 16 game week the same, but just weights the upper end more heavily. 14 game weeks lose 50 points, however. That can easily be countered by having the bonus for those weeks be 30-40-50-60-70. I like that, actually, since it's a lot harder to pick 10+ games right in a 14 game week than it is for 16. (See below.)

Another idea might be to start bonus points after 11 correct in a 16 game week. The rationale for doing so might be found in looking at the probability of getting bonus points by just randomly selecting games. In a 16 game week, flipping a coin for each game gives you a 22.72% chance of getting 10+ games. There's a 10.51% chance of getting 11+ games. For a 14 game week, there's a 9.08% chance for 10+ games. By bumping up the number required for a 16 game week, the probabilities of getting into the bonus (by a coin flip, which serves as a base line) are roughly equal (within 1.5% being "roughly").
 
2Chuck
Donor
ID: 169212110
Mon, Feb 02, 2004, 13:07 		Warning: I'm not a big statictician, just someone who is looking at it more from the "masses" perspective. I didn't even think out the logistics, but just want to throw out a few possibilities.

1. I wonder if some sort of "game of the week" option could be added. I'm not quite sure how it would work, but you could designate a game of the week (or use a standard such as the Monday night game). That game would be worth automatic double points if picked correctly, but have no extra risk on the downside. It might be a way to bring a bit more strategy into the game.

2. I know that negatives are part of the game, but there is something in human psychology that loves high scoring, even if one is no more in the game with big numbers than they were with negative numbers. I think part of the lag in pickoff is that people feel they are way out of it when they have a negative score, and stop picking. This year was a rare year where the full-season came down to the final game; I wonder how this can become the norm in the future.

3. Personally, I'd love to see a scale of:
5-10-20-40-80-150-300. It would make the game a bit different, but it would let people realize that they can make up a larger deficit in a short amount of time.

In actuality, I think the game is great overall. I just think the fact that interest drops as the season goes on is a flaw. As the season goes on, the interest should be increasing, and we should work to change that trend.
 
3Sludge
ID: 150192911
Mon, Feb 02, 2004, 13:32 		I think the more bonus points you award, the more you encourage the picking of favorites trying to shoot the moon. Rather than carefully weighing the risk versus reward of picking a slate of games, the optimal/preferred strategy then becomes picking who you think has the best chance of winning each game, which whether intended or not is not what this game was ever about. If I get a couple of heavy favorites wrong, who cares? The massive bonus points will make up for it. It's not "Pickoff" any longer, it's just "Pick 'Em". To me "Pickoff" has never been about picking who you think will win. My slates constantly include a substatantial number of upsets. Individually, I don't think that the upset will happen. Collectively, however, I bank on a few of them happening.

If a massive amount of possible bonus points are added, however, I would prefer to see something like Chuck's distribution where you really have to get into the rarified air to see the big payoff.

I also think that you don't have to worry about people being able to make up large deficits in a short amount of time. That already happens with regularity.
 
4Sludge
ID: 150192911
Mon, Feb 02, 2004, 13:37 		Hmm... now here's an idea way out of left field.

Why not make the bonus points dependent somehow on the frequency with which favorites are selected? The more heavily favorites are picked, the lower the bonus points get.
 
5JeffG
Sustainer
ID: 1584348
Mon, Feb 02, 2004, 14:04 		The escalating bonus points would just encourage more people to sit with the heavy favorites since the 70+ points they would get with the lesser picked team may now be overshadowed by the 20+ points plus bonus they could get with the more picked team. Not that this is a bad thing.

To try to encourage those who lose interest as the weeks progress, or to try to encourage more participation in the post season, perhaps "award" each of the 17 weekly winners with a berth into a post season contest where the one who does the best in the post season among those winners gets a small prize. Depending on how much extra bookkeeping you want to do, maybe add an additional berth to the person who does the best on their 'doubled picks' throught the season, the one who does the best on the Monday night games only (not counting it as double if they use it in that game, for this extra contest), and the one who does the best on Sunday night games only. People would be more apt to stay in if they had some shot at something.

Regardless, this contest is one of the things I look forward to each season, and am greatful for your continued efforts.
 
6Nuclear Gophers
ID: 230202914
Mon, Feb 02, 2004, 14:06 		Make the playoffs a week 18 performance so that bonus points can be accumulated. I think there are 11 games so have bonus points in conjunction with the # of games played. Just my 2 cents.
 
7Chuck
ID: 571132017
Mon, Feb 02, 2004, 15:43 		Guru- What percentage of the time did favorites win this year? Obviously that will vary from year to year, but it should be a good scope. I think adding the major bonus points forces people to choose: Go with lots of favorites and try to rack up bonus points (which will not always work), or pick strategic underdogs. Some weeks one strategy will win, and other wekks it will be just the opposite.

I also think that you don't have to worry about people being able to make up large deficits in a short amount of time. That already happens with regularity.

I agree here, but I think there might be a perception that early deficits cannot be overcome. Just seeing someone pull off a 700 point week might change that perception.

To try to encourage those who lose interest as the weeks progress, or to try to encourage more participation in the post season, perhaps "award" each of the 17 weekly winners with a berth into a post season contest where the one who does the best in the post season among those winners gets a small prize.

JeffG- Awesome idea! I love it. In the case of the same person winning 2 weeks, you could roll down to the #2 person in that week. In the very least, that should keep people into it through the regular season.

I think part of the reason people cut off in the postseason is that they get out of routine. For me, I will finalize my fantasy line-ups on Saturday night or Sunday morning. When I'm doing that, I also do my Pickoff picks. In the playoffs, there are no rosters to adjust, and I think pickoff slips the mind at times. I know I totally brainfarted on the Super Bowl.
 
8Sludge
ID: 150192911
Mon, Feb 02, 2004, 15:45 		The escalating bonus points would just encourage more people to sit with the heavy favorites since the 70+ points they would get with the lesser picked team may now be overshadowed by the 20+ points plus bonus they could get with the more picked team.

I disagree. That would only happen if they get themselves into a position to get 13-14+ correct. Otherwise, the bonus points would reward them less than it currently does, so you can't bank on getting 50+20 for that heavy favorite as you might now. Unless you think you can get 13-14 right every 16 game week, you'd have to bank on something less than 50 as a bonus.
 
9Sludge
ID: 150192911
Mon, Feb 02, 2004, 15:57 		I agree here, but I think there might be a perception that early deficits cannot be overcome. Just seeing someone pull off a 700 point week might change that perception.

Well, with the bonus points that you proposed, going 16 for 16 would be 605 PLUS the points for the games, which could easily lead to 1200-1300-1400 or even higher weeks, much less 700. Going 15 for 16 would be 305 PLUS the points for the games which again would be substantially higher than 700. In other words, it would be possible to get close to (possibly over) 50% of what the winning TOTAL for the year was this past season.
 
10wolfer
Donor
ID: 18639422
Mon, Feb 02, 2004, 22:10 		With tiebreakers, I would reward consistancy. If there is a tie, I would give the advantage to the person who has the most weeks with positive points.

 
11Guru
ID: 330592710
Sun, Aug 08, 2004, 13:53 		BUTT.

I'll be updating Pickoff later this month, so this is your last chance to get new ideas for 2004 on the table prior to launch.
 
12Guru
ID: 330592710
Sun, Aug 29, 2004, 20:43 		After some reflection, I'm wondering whether the best solution for bonus points is to simply eliminate them. I'm not convinced that they really add anything of value, and they do potentially distort the underlying premise of the game.
 
13Ender
ID: 0442215
Sun, Aug 29, 2004, 21:40 		That sounds good to me actually. The bonus is inherent. You get 11 right, you get points for 11 games and so on.

I like the Game of the Week idea.

 
14Sludge
ID: 24914721
Sun, Aug 29, 2004, 22:40 		I wholeheartedly second that idea, Guru.
 
15MadDOG
ID: 237512921
Sun, Aug 29, 2004, 22:57 		If that's the case then being able to "Double" a game must distort the underlying premise also.

Shouldn't that be eliminated for the same reason as the bonus points?

There is thousands of Pickems games on the internet. Is that what you want to turn the Guru version into? I say leave it the way it is. It's a unique pickem game. The doubling and the bonus points make it fun.
 
16MadDOG
ID: 237512921
Sun, Aug 29, 2004, 23:02 		A "Game of the Week" might add some fun to it also.

I short I like the variations. All of these things still take skill so I dont' see what the problem is.

It takes skill to pick a game to double, it takes skill to get 10 games right etc.
 
17MadDOG
ID: 237512921
Sun, Aug 29, 2004, 23:09 		How about if you get the "Game of the Week" correct you get your "worst" incorrect pick wiped off your score excluding the game that you doubled.

That would be cool. Might take some programming for that though.
 
18Sludge
ID: 24914721
Sun, Aug 29, 2004, 23:27 		There is thousands of Pickems games on the internet. Is that what you want to turn the Guru version into? I say leave it the way it is. It's a unique pickem game. The doubling and the bonus points make it fun.

Actually, the doubling doesn't detract from the original premise of the game (i.e. weighing risk vs. reward), while including the bonus points does. As the amount of bonus points available increases, the optimal strategy gets closer and closer to just picking the team that you think has the best chance to win, which is how most of the plain pick 'em games out there are like. By removing the bonus points, the optimal strategy actually gets as far away from picking who you think has the best chance to win as possible. It then becomes purely weighing risk vs. reward (i.e. if a team has been picked 20% of the time, the optimal choice is to pick them only if you feel they have a greater than 20% chance to win).

There are several threads that are probably still around where the optimal pickoff strategy is discussed (mostly between Madman and myself).
 
19TB
Leader
ID: 031811922
Sun, Aug 29, 2004, 23:44 		I like the reward for picking games correctly. There has to be an incentive to not pick every huge underdog.
 
20Perm Dude
ID: 2343587
Sun, Aug 29, 2004, 23:49 		There is a disincentive: If people started picking them, the odds change and they become less attractive. Also (as I reminded myself several times last season): If one team is and 85-15 favorite for example and you really do think the favorite will win without a doubt, you might as well take the 15 points and the W.
 
21Sludge
ID: 24914721
Sun, Aug 29, 2004, 23:52 		Say what, TB?

If a huge underdog has only a 10% chance of winning, then you WANT 20-30% of the pickers going for the big upset. That kind of margin makes me foam at the mouth.

It's very simple, really. Without bonus points, picking every upset (big or otherwise) would be a poor strategy, just as picking every favorite would also be a poor strategy. With bonus points, picking every favorite becomes a viable strategy. (And by "favorite", I mean picking the team you think will win whether Vegas thinks it'd be an upset or not.)
 
22TB
Leader
ID: 031811922
Sun, Aug 29, 2004, 23:59 		Look, I know you are the genius with the game, Sludge. I just play it for fun. If I see that 90& of the people are picking New England to beat Arizona, heck yeah I take a flyer on Arizona. Probably will even double them. Maybe it's not sound strategy, but it's the way I get my kicks playing the game. Now, if I am actually in contention and know that if I make some solid weekly picks I will get bonus points, maybe I take the team I expect to win the game.

You don't have to break it down for me. Trust me, I just play it for fun and all the mathematical formulas aren't going to change my mind. Guru asked for opinions and I gave mine. If it is a simple-minded opinion, just be happy you live in a sunnier world than poor TB. ;)
 
23Sludge
ID: 24914721
Mon, Aug 30, 2004, 00:03 		Guru -

Just for kicks (since not having them would likely have changed the way people picked), who would have won last year if bonus points were removed?
 
24TB
Leader
ID: 031811922
Mon, Aug 30, 2004, 00:06 		Can I guess? I bet I know.....lol
 
25MadDOG
ID: 237512921
Mon, Aug 30, 2004, 00:22 		Ok, if the bonus points reward people for picking favorites then why not add another game you can "double" or even a game that you can "triple".

Let people get extra points that way.

And I still like that idea of eliminating your worst score if you hit the game of the week.


I guess eliminate the bounus points but add other ways to get more points.
 
26Sludge
ID: 24914721
Mon, Aug 30, 2004, 00:25 		Well, of the top 10, here's the breakdown of bonus points awarded throughout the season:


Team	Bon	BonPts	Score	Diff
DR	27	1350	2906	1556
Vilica	25	1250	2807	1557
Mallon	17	850	2741	1891
Sludge	13	650	2680	2030
Rabid	16	800	2493	1693
Beer	24	1200	2450	1250
Eric	26	1300	2312	1012
yesno	19	950	2293	1343
Scooby	13	650	2184	1534
bwurst	22	1100	2161	1061


So, yes, it looks like I would have won.

I have had no problems making my strategies public in the past (I've even made the software I wrote and use publicly available), and I'll do the same here. No sweat off my back.

I want the bonuses eliminated because it gives me a better chance to win (do not, in any way, interpret this as me saying that it makes me a lock to win; it doesn't). Really, it's as simple as that. I'm not being altruistic here, I'm looking out for number one. But I do believe that a careful weighing of the risk versus reward is what Guru had in mind (feel free to correct me, Guru), and not having bonus points, as I've said before, is closest to that. Having bonus points gives people who don't want to do that weighing a viable strategy that doesn't require it or it gives people who do weigh risk vs. reward an easy tie-breaker for those games that are a toss-up.
 
27MadDOG
ID: 237512921
Mon, Aug 30, 2004, 00:30 		I doubt that the guys that beat you actually picked the favorites every week.

How do you know if anybody actually used that strategy to beat you?
 
28Perm Dude
ID: 2343587
Mon, Aug 30, 2004, 00:30 		Then how about bonus points only for picking x number of underdog winners correctly in a week?
 
29MadDOG
ID: 237512921
Mon, Aug 30, 2004, 00:36 		Make the "Game of the Week" the game with the smallest point spread. You get to eliminate your worst score if you pick it correctly or eliminate your best score if you miss that game.

That would put some serious importance on a difficult game to pick.
 
30MadDOG
ID: 237512921
Mon, Aug 30, 2004, 00:37 		That's a good idea too Perm Dude.
 
31MadDOG
ID: 237512921
Mon, Aug 30, 2004, 01:02 		Just figure out the average number of underdogs that win every week and start the bonus points around that number some where.
 
32MadDOG
ID: 237512921
Mon, Aug 30, 2004, 01:10 		Or eliminate 1/2 of your worst score if you get the "game of the week" correct and 1/2 of your best score if you miss the "game of the week".



I'm going to bed. LOL.
 
33Ender
Donor
ID: 013443221
Mon, Aug 30, 2004, 09:23 		Now that I've slept on it...

Bonus points aren't all bad. I understand the premise of the game, but I don't see how it is undermined by attempting to pick as many games correctly as you while at the same time balancing risk/reward. There is simply more than one kind of reward. If there were no bonus points why not just pick all the 20% (or 25 or 30) or lower underdogs each week?
 
34Sludge
ID: 475323018
Mon, Aug 30, 2004, 09:27 		MD - I doubt that the guys that beat you actually picked the favorites every week.

I never said that the optimal strategy is to pick who you think will win (and, again, I'm vaguely defining that as who you give the best chance to win):


As the amount of bonus points available increases, the optimal strategy gets closer and closer to just picking the team that you think has the best chance to win, which is how most of the plain pick 'em games out there are like.

With bonus points, picking every favorite becomes a viable strategy.


MD - How do you know if anybody actually used that strategy to beat you?

Have you even bothered to look at the standings of the top teams? For the most part, they picked a great majority of favorites going for the bonus, and a large number of the upsets picked were of the 40-60 variety. Of course they altered their picks some because of the points awarded being dependent on the percentage, but for the most part they were trying to maximize their chance of getting in the bonus by trying to maximize the number of picks they got right. The ones at the top of the leaderboard just happen to represent those that did a better job at picking the upsets. That sounds like a pretty good description of the winners of a run-of-the-mill non-spread pick 'em game.

PD - Then how about bonus points only for picking x number of underdog winners correctly in a week?

I'm not so sure I would like that either. That just shifts the best strategy in the other direction.
 
35Perm Dude
ID: 2343587
Mon, Aug 30, 2004, 09:37 		Sludge, I think we need to look at the intent for the bonus points, rather than the effect. IMO, bonus points for picking a large number of underdogs in a week encourages underdog picking, but doesn't take away on the other side (bonus points is a light weight on the otherside. Does anyone try to pick more winners because of bonus points being awarded for W's?

Some weeks, bonus points are extraordinarily easy to attain. Other weeks, no. The picking of underdogs (going for the risk/reward) is dampened by bonus points for wins, since it is typically the favorites who win.

If we don't get rid of bonus points, make them a function of picking successful underdogs, and make it a good high number (like 6 or even 8) to make them rare enough to really be a bonus, and also rare enough to make it a real accomplishment.
 
36Guru
ID: 330592710
Mon, Aug 30, 2004, 09:53 		Last season, picking 100% consensus favorites would have placed at approximately the 95th percentile. That is the issue that I'm trying to mitigate.

There are some interesting approaches to amend the bonus point formula suggested above, but all seem a bit complex. There is a lot to be said for simplicity - which has me leaning toward simply eliminating bonus points for this season.
 
37Sludge
ID: 54692111
Mon, Aug 30, 2004, 10:29 		The picking of underdogs (going for the risk/reward) is dampened by bonus points for wins, since it is typically the favorites who win.

I would argue with that.

Of the following two, which carries more risk?

Team A is picked by 20% but has a 15% chance to win.

Team B is picked by 85%, but has an 80% chance to win.

Based on how I posed the question, the answer should be obvious: They both carry the same risk in the long run.
 
38Sludge
ID: 54692111
Mon, Aug 30, 2004, 10:35 		Ender - If there were no bonus points why not just pick all the 20% (or 25 or 30) or lower underdogs each week?

By all means, please do.
 
39Motley Crue
Leader
ID: 439372011
Mon, Aug 30, 2004, 10:51 		Guru - Last season, picking 100% consensus favorites would have placed at approximately the 95th percentile.

How about only allowing doubling on underdogs?

How about giving a 10 point bonus to each underdog chosen, and/or a -10 point penalty to each favorite selected?

How would that affect things, Sludge?
 
40MadDOG
ID: 46739309
Mon, Aug 30, 2004, 10:51 		I'm gonna say it again. Guru, please don't turn this game into just another Pickem game.

The variations are what make it a good game.

There really is only one guy complaining about the bonus points. The bonus for picking underdogs sounds like a very good idea but you can't make it kick in at 6 or 8 underdogs. I think it's very rare when 6 underdogs win in a given week. 4 underdogs seems like a good spot to start the bonuses.
 
41Ender
Donor
ID: 013443221
Mon, Aug 30, 2004, 11:02 		Sludge, you don't own a monopoly on exagerating for effect :) I never said it was the perfect strategy, but I would argue that eliminating the bonus points moves it "closer" to simply picking the underdog anytime there is a sub 20 underdog. I know sometimes you the probability doesn't warrant that and you should just grab the 15 points, but you know what I am talking about.

Guru, I am actually inclined to agree a bit with MadDOG here. You raised the question and there appears to be only one person in favor of changing the game in that regard and he confesses it is only to serve his own purposes. It doesn't make the game any better per se, but it helps a single player win. I would never see that as a reason to change a rule.
 
42Sludge
ID: 54692111
Mon, Aug 30, 2004, 11:02 		Re: #40

Guru - Last season, picking 100% consensus favorites would have placed at approximately the 95th percentile. That is the issue that I'm trying to mitigate.

It already IS "just another Pickem game"! I rest my case.

MC - How would that affect things, Sludge?

Without bonus points, it would mean that you would only pick a favorite if the percentage picked is 10 under what you think the probability of them winning is. With bonus points, that would translate to swinging the needle back towards a careful weighing of the risk vs. reward, possibly even past it in the direction of picking upsets. I can't say exactly how much, though.
 
43TB
Leader
ID: 31811922
Mon, Aug 30, 2004, 11:26 		It is not just another pick'em game. In a normal pick'em game, whoever gets the most games correct, wins. In this game, you can pick 5 correct winners and score more points than the guy who picked 9 correct and even sometimes 10. The whole point of any pick'em game is to pick as many games correct as possible. The risk for picking an underdog in this game is to lose out on bonus points if you are wrong. The reward is getting lots of points if you are correct.
 
44MadDOG
ID: 46739309
Mon, Aug 30, 2004, 11:44 		PD - Then how about bonus points only for picking x number of underdog winners correctly in a week?

I'm not so sure I would like that either. That just shifts the best strategy in the other direction.


Are you saying that if Guru gave out bonus points for picking underdogs, that picking all the underdogs is a viable strategy?

I doubt that. You would lose way too many points on all the missed games where the favorite won. You couldn't overcome the lost points just with bonus points. Matter of fact, I don't think you would even be in the top 100 with that strategy.
 
45Sludge
ID: 54692111
Mon, Aug 30, 2004, 12:08 		MD - Are you saying that if Guru gave out bonus points for picking underdogs, that picking all the underdogs is a viable strategy?

Depending on where the bonus would start, sure, but it wouldn't be the "best". I chose the word "viable" very carefully. "Viable" is not the same as "optimal" or "best".

Let me put it this way. Consider the following game:

Team A 25% - Team B 75%

Suppose that you feel that Team B has a 74% chance of winning. In the expected value, that's a 1 point margin in favor of A. If no bonus points are awarded, your optimal decision is to select Team A. If bonus points are awarded simply for getting games right, then the optimal decision is to select the favorite, B. The more bonus points awarded for getting games right, the larger that margin can be in favor of A while still having B be the optimal selection.

That's what I mean when I say that it is an easy tie-breaker for toss-ups, and what I mean when I say that the best strategy would shift. The reverse is also true. If you award bonus points for getting X underdogs right, then it would also serve as an easy tie-breaker for toss-ups (i.e. it would represent a shifting in the best strategy towards the underdogs).
 
46Sludge
ID: 54692111
Mon, Aug 30, 2004, 12:18 		TB - It is not just another pick'em game. In a normal pick'em game, whoever gets the most games correct, wins. In this game, you can pick 5 correct winners and score more points than the guy who picked 9 correct and even sometimes 10.

Well, sure you can. But if that's such an attractive prospect to you, TB, then why are you arguing against getting rid of the bonus points? Surely you realize that awarding bonus points for getting games correct makes the above scenario less likely? (No, not impossible or even rare. Just less likely.)
 
47Perm Dude
ID: 2343587
Mon, Aug 30, 2004, 12:21 		I never said picking all underdogs is any kind of strategy at all. I do believe that awarding bonus points only for picking a certain number of underdogs correctly makes underdogs slightly more attractive, adding to the "reward" while making little change in the "risk."

Which is why bonus points for all Ws favors the favorites right now. The reward is given for any W, and typically the easiest way to get the W is to pick the favorite.

And Sludge, we are not looking at the long run, since bonus points accumulate only for discrete periods (each week). Eliminating the bonus points for favorites doesn't change the risk for picking the favorite, but it makes underdogs oh so more attractive.

This all might be moot, however, given the programming difficulties Guru points out might be incurred.

pd
 
48Sludge
ID: 54692111
Mon, Aug 30, 2004, 12:32 		Ender - Sludge, you don't own a monopoly on exagerating for effect :)

I wasn't exaggerating. Being a bit flippant? Yes. Exaggerating? No.

Ender - I know sometimes you the probability doesn't warrant that and you should just grab the 15 points, but you know what I am talking about.

You'd be surprised how often it is. It's a bit more than "sometimes". It is my experience that folks playing this game tend to slightly overpick the favorite. Why is that? In the first year, you might chalk it up to people being unfamiliar with the underlying principles, but we never got a chance to see if the collective would, after more familiarity, adjust to this. In subsequent years, the reason for it is simple: the availability of bonus points (did bonus points start in year 2, Guru?). I think you're basing your thoughts on what would happen if the bonus points were eliminated on the behavior of the past couple of years when bonus points were in effect.
 
49MadDOG
ID: 46739309
Mon, Aug 30, 2004, 12:33 		I would like for Sludge to prove to me that picking all underdogs is a viable strategy if Bonus points start with 4 underdogs.


Sludge, you're making alot of claims but you have nothing to back it up. I doubt very much that you can even come close to winning this game by picking all underdogs when the bonus points start at 4 underdogs.

Alot of these games come close to a 50/50 split on people picking them. And you're gonna lose close to 50 points for every game like that. How many games per week have a Vegas point spread of 3 or less? The answer is alot. And these are the games you are gonna lose alot of points when the underdog doesn't win. There is no way you will overcome that with bonus points.
 
50MadDOG
ID: 46739309
Mon, Aug 30, 2004, 12:41 		And since picking favorites IS a viable strategy to you because of the bonus points that means that 10 or more favorites are winning every week because you have to have 10 winners to get bonus points.

How are you going to overcome more than 10 losses every week with bonus points starting at 4 underdogs? The answer is you won't.
 
51Guru
ID: 330592710
Mon, Aug 30, 2004, 13:01 		Here's my thinking:

Suppose there are only four games, and all are 75/25.

Person A picks all favorites, and 3 win. He earns 25 for each favorite win, and loses 75 for the loss. The net score is zero.

Person B picks all underdogs. He gets 75 for the one that won, but loses 25 on each of the losses. The net score is zero.

So far, both players have a score of zero. Now, since Player A got 3 games right, does he deserve to be ranked ahead of Player B (via some sort of bonus scheme, tiebreaker, or whatever)?

If you think the player with 3 correct picks deserves a better result, then it seems that you want this to be more like a traditional pickem game. If you you say that no one should be ranked higher, then you are arguing that the fundamantal scoring process should not be biased in favor of picking winners, but only biased in favor of picking those who "outpicked" the odds. In this case, a tie seems appropriate.

In an efficient market, the odds should actually reflect the bonus system. If a system is biased in favor of picking winners, then the odds for favorites should be slightly stronger, reflecting the higher bonus point opportunity. The results in 2003 don't seem to reflect this.

However, I looked up the results from the past 4 years, all which used the same bonus formula. Although last year produced a 95%-ile result for picking all favorites, the other 3 years were not so kind to that strategy:

Year &-ile
2000: 37%-ile
2001: 82%-ile
2002: 32%-ile
2003: 95%-ile

So while I am reacting to the 2003 results, those do not appear to be the norm over the last 4 years. On that basis, perhaps the problem is not persistent, and a cure is not really needed.

However, I do think that awarding a bonus for getting 10/16 is probably too generous, which gets me back to a formula adjustment. For 16 game weeks, more than 30% of entries earned a bonus last year.

A simple adjustment would be to raise the bonus hurdle to 12 correct picks for a 16 game week. Using last year's results, only 10% of entries would have earned bonuses. We could leave the hurdle at 10 correct picks for 14 game weeks. Only 13% earned a bonus during those weeks last year.

In summary, I'm not persuaded that eliminating the bonus formula makes the game more like other pickem games. In theory, I think the reverse is true. But thus far, there seem to be more people (in a very limited sample, I realize) who would like to retain the bonus system. How do you all react to this potential adjustment (i.e., raising the hurdle to 12 correct picks for 16 game weeks)?
 
52MadDOG
ID: 46739309
Mon, Aug 30, 2004, 13:11 		Sounds good to me. Very rarely are you gonna get to 12 wins without picking 3 or 4 underdogs.


Guru, what about adding one more gimmick? Say one more double game. Or a triple game? Or a game of the week with some sort of bonus?
 
53MadDOG
ID: 46739309
Mon, Aug 30, 2004, 13:15 		If you have a "triple game", people might try to put that on a favorite ,like a team picked by 75% to gain 75 points , where if they miss the results could be disastrous. (-225 points)
 
54MadDOG
ID: 46739309
Mon, Aug 30, 2004, 13:40 		I just checked the "Monday Night Football" schedule and there is alot of very close games on that schedule.

Some sort of bonus related to the Monday Night game would add some excitement. Maybe picking the total score and getting a bonus if you can come within 3,4,5 points of the total score. 50 points if you can come within 5 points of the total score with no penalty?




 
55MadDOG
ID: 46739309
Mon, Aug 30, 2004, 13:43 		Or just make the Monday Night game an "automatic double" like somebody has already suggested.
 
56Sludge
ID: 54692111
Mon, Aug 30, 2004, 14:20 		Guru - How do you all react to this potential adjustment (i.e., raising the hurdle to 12 correct picks for 16 game weeks)?

My reaction should be fairly predictible, especially given my post #1. Anything that moves away from bonus points is fine with me.
 
57MadDOG
ID: 57203013
Mon, Aug 30, 2004, 14:23 		I really like the idea (mine- LOL) of if you hit the "Monday Night" game you get your worst score from the previous games eliminated. In effect you get a bonus in the amount of your worst score. Conversely if you miss the "Monday Night" game, you lose points equal to you best score from the previous games. In effect a penalty in the amount of your best score.

This would be on top of the regular points you get/lose for the "Monday Night" game.

It would be sort of a variable bonus. You might have to exclude the "doubled game" from this. That would really be cool if you could pull it off.

Doesn't seem like it would be that hard to program.
 
58Cuc
ID: 557223013
Mon, Aug 30, 2004, 14:51 		I'd like to see bonus points that go up progressively starting at 12 wins during a 16 game week.

50 bonus points for the 12th win
100 for the 13th win
150 for the 14th
200 for the 15th
250 for the 16th.

This would be (your score)+50+100+150+200+250

This would keep people interested because anyone could vault into 1st place by picking a perfect slate (which rarely happens, but it's something to shoot for).

This probably undermines the original purpose of the game, but it's just a thought.

 
59Cuc
ID: 557223013
Mon, Aug 30, 2004, 14:52 		I like the "Monday Night Game of the Week" idea.
 
60nate686
ID: 467323019
Mon, Aug 30, 2004, 20:32 		I love your pick'em Rotoguru, it is original and a very fun game to play. In my opinion, you shouldn't change a single thing about it, this pickoff is the best game goin on the Net today in terms of pick'ems. By eliminating bonus points altogether, then we simply have a pick'em mainly as upset picks will not be nearly as valuable and increasing the # of wins needed for bonus points will more than likely make more people pick the favourites because the risk/reward for picking an underdog may not allow the contestant to reach 12 victories. 10 was much more attainable, and I can tell you from someone who has not had great success in this game that it is difficult to pick in this game. Just look at my team, The Knuckles, I had a great week 8 or 9 I believe and I completely tailed off after. The %'s mixed with the bonus points MAKE you consider picking the underdog because you must analyze risk/reward. If there are no bonus points or the bonus points are difficult to achieve, I believe the masses will not pick many underdogs because the risk doesn't meet the reward. I don't have any stats to back me up, but that's just my opinion.
 
61DR Stars
ID: 46732106
Wed, Sep 01, 2004, 23:18 		I know I'm late to the party, but here's my opinion:
First off I think elevating the 10 pick threshold is a good idea, going to within 4 of the total is a more realistic "bonus" than 10/16.

Adding a special game like the MNF or a triple game option sounds less attractive to me, it would put too much importance in one game where we should be looking at a more balanced outcome. IMO.

And for Sludge in post #26, you say you'd have won if the rules were different, but maybe the rest of us would have used a different strategy in picking games if that were the case, so no one knows the answer to that. ;)
 
62Sludge
ID: 475323018
Wed, Sep 01, 2004, 23:45 		DR -

Yeah, that's why I said the following in post 23:

"Just for kicks (since not having them would likely have changed the way people picked), who would have won last year if bonus points were removed?"
 
63Sludge
ID: 475323018
Thu, Sep 02, 2004, 00:36 		And I would point out that you make my point for me. What "different strategy in picking games" would have been employed? It certainly wouldn't be relying even more heavily on the favorites, that's for sure.

Just so we're clear, that's not pointed at you DR, as it's apparent that you, at least in part, agree with me. I also think you're a class act.

Note to MadDOG: "in part" means that it looks like he agrees with some fraction between 0% and 100%, not inclusive. I thought I would make that clear since you seem to like to take things people say and automatically jerk their words and positions to an extreme. And as to the ridicule you tossed at me in posts 49 and 50, I would suggest doing a bit of homework yourself before posting things like that. Take a look back at the standings for the year before last. Even without bonus points, picking all consensus underdogs was a mere 80 points out of 50th place. With the most conservative possible assignment of the 101 underdog wins, a minimum of 1950 points would have been gained. Take away bonus points for the favorites, and there's no telling how high it would have been ranked. (I can't figure it up for sure without access to the week-by-week results.) So much for "I don't think you would even be in the top 100 with that strategy."

Just to clear up something that's been nagging me, I never said or implied that picking favorites would place you near the top on a consistent basis. I've only ever been concerned with the effect on picking strategies that the bonus points has. As I have demonstrated rather nicely in the two years prior to last year, even the "optimal" (as close as I can get it) strategy can suck eggs some years.
 
64Motley Crue
ID: 181650
Sun, Sep 05, 2004, 15:11 		Guru,

I can't make any picks in Pickoff right now.

It says it's too late to lock in picks.

//MC
 
65Perm Dude
ID: 2343587
Sun, Sep 05, 2004, 15:51 		Ditto.

BTW, I set up an open division if anyone wants to put a team in there, "Perm Dude's Open Division."
 
66Guru
ID: 330592710
Sun, Sep 05, 2004, 17:19 		Oops - the system is set up tp automatically lock out all games at 1pm on Sunday. I forgot to tell it that today doesn't count.
 
67MadDOG
ID: 14820279
Mon, Sep 27, 2004, 14:13 		Looks like raising the bonus points to 12 correct picks has made the game more competitive.

The top guy only has 869 points.

I'm negative but I don't feel like I'm out of it at this point.
 
68Motley Crue
Leader
ID: 439372011
Mon, Sep 27, 2004, 14:34 		You are, because you can't get any bonus points to catch up!!

I have doubled a loser 3 weeks in a row now. I may just stop doubling altogether, since I am losing alot of points because of it (doubled Miami last night). I have picked pretty well otherwise.
 
69MadDOG
ID: 14820279
Mon, Sep 27, 2004, 14:37 		You can still have big weeks even without bonus points. And you can still get bonus points at 12.

A couple big weeks in a row and your right back in it.
 
70Motley Crue
Leader
ID: 439372011
Mon, Sep 27, 2004, 14:38 		I was just messing with you. If you can consistently pick well, you'll be right back in it before the season's over.
 
71MadDOG
ID: 14820279
Mon, Sep 27, 2004, 14:42 		You're always messing with me. I'm gonna find out where you live and punch you in the face.

:)
 
72Motley Crue
Leader
ID: 439372011
Mon, Sep 27, 2004, 14:44 		I live at TB's house. And I look just like him. Come get me.
 
73MadDOG
ID: 14820279
Mon, Sep 27, 2004, 14:46 		Hmmmmm you and TB are the same person.

I did not know that. Why the double identity?
 
74TB
Leader
ID: 031811922
Mon, Sep 27, 2004, 20:48 		Here is our secret address

If I could go back in time, I would have waited to make pick-off teams until this week. Without going to look, I think they are all deeply negative.

Is there a prize for the worst loser? If so, I know all my teams are in the running.
 
75Motley Crue
ID: 181650
Mon, Sep 27, 2004, 22:11 		LOL!

How you like them apples, Jerry?!
 
76Perm Dude
ID: 2343587
Tue, Sep 28, 2004, 10:50 		After 3 weeks, I've picked 25 winners and have 31 points!

I shouldn't fret--at least I'm still positive.
 
77Motley Crue
Leader
ID: 439372011
Wed, Dec 29, 2004, 15:14 		Guru, something strange is going on when I go into Pickoff to make my picks. The percentages are there in blue, but there are other numbers and percentages in black as well.



I can't see that picture. Not sure if it's because I'm at work or something else.

No problem making my picks. I just wanted to let you know that it looks different, in case something needs to be fixed.
 
78Guru
ID: 330592710
Wed, Dec 29, 2004, 18:11 		Should be OK now.
 
79MadDOG
ID: 41143108
Wed, Dec 29, 2004, 18:44 		I think you should add the point spread to Pickoff. That would make it really interesting.


Or make a point spread version of the game. Like you could have one login name but have a button where you could switch between the regular version and the point spread version. Each guy could play both games at the same time.
 
80MadDOG
ID: 41143108
Wed, Dec 29, 2004, 18:46 		You don't have to give out prizes for the point spread version. Just make it an option for people to compete in just for fun.
 
81The Dan's Your Daddy
Donor
ID: 022792222
Mon, Jan 24, 2005, 15:14 		I agree. It would raise participation in our Best Bets threads and attract the most serious of the sports fanatics: those who do it for a living!

I pledge $25 for the establishment of this option!
 
82Guru
ID: 330592710
Wed, Aug 31, 2005, 12:34 		I'm getting a late start at reviewing potential scoring change options for the 2005 season.

Consequently, I don't have time at this late date to add a point spread option. Perhaps I can dabble with this idea and make it available on a trial basis for part of the season - although I make no commitment for that right now.

I could tweak the bonus formula again, however. I just read through last year's comments on this topic, and I am still not persuaded that the bonus formula is necessary as a way to distinguish this game from other pick 'em games. I think the underlying game is already distinctive.

I also see that a straight slate of picking consensus favorites last year produced approximately an 85%-ile result. Of the total 682 points earned for that strategy, 500 were from bonus points.

However, given last year's feedback that bonus points seem to be a "desirable" feature, perhaps the best approach for this year is to again move incrementally to a level which makes bonus points less influential.

For example, last year, bonus points kicked in for the 12th correct pick in a 16 game week, and with the 10th correct pick in a 14 game week. If I raised the bar by one correct pick (i.e., awarding the bonus starting with the 13th or 11th correct pick), the consensus favorite strategy would have only earned 150 bonus points.

I'm going to do a bit more dabbling with this idea, but I thought I'd throw it out there, just in case there is an unexpected reaction.
 
83Guru
ID: 330592710
Wed, Aug 31, 2005, 12:59 		Just to throw some more numbers out there...

Last year, the average team earned 140 bonus points over the year.

If we had used a bonus point scheme which required one more correct pick (as proposed in the post above), the average team would have earned a bonus of about 46 points over the whole season. (I realize that the adjusted bonus formula might have led to some changes in picks, but I suspect the differences would not have been significant.)

Furthermore, if we had used the proposed bonus formula, the average total score for the season would have been -5. I suspect this is a bit skewed, since it includes some entries that were abandoned over the course of the season, and those probably had worse than average scores. So the average score of the persisting entrants was probably somewhat better.


 
84Motley Crue
Dude
ID: 439372011
Wed, Aug 31, 2005, 13:53 		Ah, how funny it is to read MadDOG's posts once again.
 
85 Nadim
ID: 5522212
Fri, Sep 02, 2005, 08:28 		If Bonus is included then, for games that are in the 55%-45% or similar range, one will pick the favorite since the -10 or so extra points are compensated for by the big gain in bonuses in case the favorite wins. This takes the fun out of these close games.

IMHO, the pickoff game would be ideal if there was no Bonus. This makes picking the games (especially the close ones percentage wise) more challenging and more fun.

Nadim :-)
 
86youngroman
ID: 298482214
Fri, Sep 02, 2005, 08:51 		I thought of this last year, but didn't post it here: what if we not only can pick the winner of the 14/16 games per week but also some head-to-head battles between players, lets say:
who throws more yards this week: Peyton Manning or Trent Green?
who rushes more: Priest or LT?
who creates more turnovers: BUF or BAL?

this could be a seperated contest to the normal pickoff game but it may be fun to play.
 
87 Rick Zurbrick
ID: 3183416
Sun, Sep 04, 2005, 17:03 		There are way too many possibilites to include any sort of a head to head match up. At least IMHO
 
88Guru
ID: 330592710
Sun, Sep 04, 2005, 20:21 		Football Pickoff has been launched for 2005.

For now, there are no rules changes from last year, other than the bonus formula adjustment described in post2 82-83.

Please let me know if you encounter any apparent glitches in game operations.

I have already adjusted the week 2 schedule to show N.O at NYG, rather than vice versa, as the game will apparently be moved to New York. If it is played on Saturday, the usual freeze rules will apply for early games.

 
90Species
      Leader
      ID: 07724916
      Tue, Sep 13, 2005, 13:30 		I have an idea: all of the rest of us just MAIL IT IN and declare Sludge the winner. Fxxxing statisticians! ;-) I can't compete with his statistical modeling!!
 
91Sludge
      ID: 27751510
      Tue, Sep 13, 2005, 13:38 		Bah. I'll come back to earth. If there's one thing that this game should have taught everyone by now is that there's no such thing as an insurmountable lead (the last couple of weeks excluded). This week was pure luck... an outlier. I've said it before, and I'll say it again. My method isn't geared towards instant results. It's about piling up consistent points over the long-run by figuring out which teams have been over-picked... or trying to anyway.
 
92Species
      Leader
      ID: 07724916
      Tue, Sep 13, 2005, 13:49 		I hear you Sludge. Well, my strategy this year was to put 100 entries in, have them pick ALL of the Vegas favorites early in the week, then run an algorithm to switch ALL of those picks 2 minutes before kickoff, effectively screwing YOU......

....allowing my legitimate picks to cruise to the finish line! ;-)
 
93Sludge
      ID: 27751510
      Tue, Sep 13, 2005, 14:00 		That strategy's been there for the taking since I told everyone exactly how I pick. Of course, I'm guessing Guru wouldn't take too kindly to it. But I don't think forming a bloc would be...
 
94Guru
      ID: 330592710
      Sat, Sep 17, 2005, 17:30 		Just a heads up - I will be away on Sunday, Sept. 18 from before noon until evening (8pm or so EDT), so I won't be able to do any game administration (update points, troubleshoot, etc.) during that period. I usually try to post scores from the early games shortly after they end, but that won't be the case tomorrow.