Let's Put this Trend Crap to the Test

kroyrunner89 · 09-30-09, 01:14 AM

This week's plays according to the trend:

Chicago -10
New Orleans -7

Let me just say really quick that I'm not backing these plays, anyone who wants to see my plays can see my weekly thread as usual. This will not count towards my record.

bombCanada · 09-30-09, 07:42 AM

Originally posted by kroyrunner89

I'm sick of hearing about trends. Anyone can manipulate data to find positive results going back, whether they're manipulating what stats you look at, the time frame, the required point spread of the team you should be betting... etc. For every trend you find supporting one team in a given week you can find one supporting the other team. It's all meaningless in the end. So, let's put this to the test. I'm going to test this 31-6 ATS trend saying you should bet favorites of 6.5 or more before their bye week. I'll track this the rest of the year and we'll see if it hits at a similar clip. I know that one trend doesn't prove or disprove whether trends work or not, but hopefully this'll open some eyes. I don't care whether you agree or disagree with me, let's see what the results are when this is all said and done.

I absolutely agree with you on this, Kroy. The data set is 37 games, which for statistical purposes is a really small number... the degree of certainty is too low.

I'm left asking "where did the -6.5 number come from?" and "why are only big favorites motivated by the desire to think for the next two weeks about a winner not a loser?" And "aren't all teams motivated by the desire to think about winners all the time?"

Very interested in seeing your results. Be prepared for the trend will continue right up until it doesn't. It could work great for the next 2 or 3 years and then stop. Or it could stop next week.

Limey · 09-30-09, 09:53 AM

I agree in general and I pay virtually nil attention to 'trends' myself. However, the one mentioned in this thread does at least have some kind of logic behind it, unlike many 'trends' which are clearly meaningless.

kroyrunner89 · 09-30-09, 11:10 AM

That's part of why I'm testing this one, because it's one of the few I've seen that actually has some decent reasoning behind it. But still, where does the number 6.5 come from, and why is it such a small sample size (I'm sure there have been many more games that would apply to this situation)? We'll probably see 7 or 8 plays this season from this trend, and although that's way too small of a sample size to prove or disprove anything I'm hoping we see this trend post a 4-4 record or worse so that people get a feel for how unreliable these really are.

Bogart45 · 09-30-09, 08:24 PM

Ha, I was going to do the same myself when I saw that post. Should you also start with last week's games, Phi -7.5 qualified and won. I think you should count it since it's all the same year.

LT Profits · 09-30-09, 10:23 PM

I think this one has more merit than saying a team is 27-5 when playing at 4:00 ET on artificial turf off of an Under if their opponent is off of three straight SU and ATS wins. LOL

Nothing wrong with examining trends that have sound logic behind them. Also, yes all teams have incentive to win before a bye, but decent sized favorites are almost always the better team, so given equal incentive, the better team usually wins barring bad luck.

As for that "magical" 6.5 number, let me start by saying that I found this trend on a different site and merely double checked to see if it was correct, which it was (30-6 at the time + Philly last week). That is since 2002. I suppose I could go back further and play with different sized favorites, but that would take a lot of time. I guess one theory is that -6.5 ensures that the favorite is really the better team? Who the fock knows?

Anyhow, I just threw the stat out there, it is not as if I use it exclusively to make plays myself. In fact, I was all set to play the Jets this week, but laid off because of this angle. However, I may actually go against this trend if the Jets hit 7.5.

kroyrunner89 · 10-01-09, 12:39 AM

Originally posted by Bogart45

Ha, I was going to do the same myself when I saw that post. Should you also start with last week's games, Phi -7.5 qualified and won. I think you should count it since it's all the same year.

Good call, didn't think about that. So it's 1-0 so far. Sorry to jump all over this one LT, not trying to attack you as a capper or anything like that. You just happened to throw one out there at a time when I was bored and looking for something to do. This one seems like the reasoning behind it may be fairly sound, but still let's track it and see what happens.

bombCanada · 10-01-09, 06:09 PM

Originally posted by kroyrunner89

Good call, didn't think about that. So it's 1-0 so far. Sorry to jump all over this one LT, not trying to attack you as a capper or anything like that. You just happened to throw one out there at a time when I was bored and looking for something to do. This one seems like the reasoning behind it may be fairly sound, but still let's track it and see what happens.

Kroy this is an excellent test of trending specifically because it's so simple. It's not what stock or future systems traders would call "curve fit", or at least not very much. There's a lot to learn from validating or invalidating this simple trend system, because if it's unreliable, then I can guarantee you that any 5 condition trend must fail.

Trends are appealing because they're so specific and so failsafe from an execution point of view. They're the triumph of emotion over reason.

LT I appreciate your good humor about this. Would you mind posting the entire data set, or a pointer to the original article ?

thanks

kroyrunner89 · 10-04-09, 08:52 PM

This trend is 3-0 so far this year, not a bad start! Let's see what happens though, still have a ways to go this season!

kroyrunner89 · 10-12-09, 02:45 AM

I forgot about this thread, oops. Okay a couple corrections here. First of all, I had NYG as a play last week for some reason, but they weren't entering a bye. Not sure where that came from. So this trend was sitting at 2-0 coming into this week. Then, Dallas would have been a play this week since they were 8 point favorites. Well, they didn't cover.

Trend after Week 5: 2-1 ATS

bombCanada · 10-29-09, 12:29 PM

Any update or thoughts on this trend crap ?

kroyrunner89 · 10-29-09, 03:54 PM

Sorry, I completely forgot about this! I'll play catch up real quick and see where things stand

kroyrunner89 · 10-29-09, 04:00 PM

Week 6: No Plays
Week 7: Patriots -14.5 WIN

After Week 7: 3-1 ATS

So doing pretty well so far, no plays again this week though. I'll try to keep my eye on this and see if there's any more plays this season

GOGOBIRD · 10-29-09, 05:12 PM

trends are crap and for people that don't know football

Indecent · 10-29-09, 08:21 PM

Originally posted by GOGOBIRD

trends are crap and for people that don't know football

And your wisdom is limited to the awareness of your own ignorance.

HoulihansTX · 10-29-09, 08:36 PM

Originally posted by GOGOBIRD

trends are crap and for people that don't know football

Naw bro really your supposed to pick your spots and bet against trends.

Serbone · 10-29-09, 09:02 PM

Originally posted by kroyrunner89

I'm sick of hearing about trends. Anyone can manipulate data to find positive results going back, whether they're manipulating what stats you look at, the time frame, the required point spread of the team you should be betting... etc. For every trend you find supporting one team in a given week you can find one supporting the other team. It's all meaningless in the end. So, let's put this to the test. I'm going to test this 31-6 ATS trend saying you should bet favorites of 6.5 or more before their bye week. I'll track this the rest of the year and we'll see if it hits at a similar clip. I know that one trend doesn't prove or disprove whether trends work or not, but hopefully this'll open some eyes. I don't care whether you agree or disagree with me, let's see what the results are when this is all said and done.

OK, but some trends matter a lot, some do NOT matter a lot.
IMO this trend is a "consideration", but would not be a significant factor in my betting decision. Depends on all the other factors.

bombCanada · 10-30-09, 08:59 AM

Originally posted by Serbone

OK, but some trends matter a lot, some do NOT matter a lot. IMO this trend is a "consideration", but would not be a significant factor in my betting decision. Depends on all the other factors.

I don't think the point of this thread is determine whether or not this particular trend matters.

We're simply tracking in real time whether this trend remains valid in the 2009 season.

The reason why we're interested in it is because it's such a simple trend (bet faves of 6.5 or more before their bye week) as compared to most other trends (involving full moons, green cheez, snow storms, road dogs, what the cheerleaders are wearing, etc) which are so horribly complicated they can't possibly be valid.

We'll have our answer after the last bye week, and then what you do with it is up to you. I don't think anyone here is risking any money on it, although Kroy might (or might not) choose to add it to his models.

Good luck!

TGoat · 10-30-09, 09:36 AM

Originally posted by kroyrunner89

That's part of why I'm testing this one, because it's one of the few I've seen that actually has some decent reasoning behind it. But still, where does the number 6.5 come from,

It's a back-fit, obviously.

Seriously, do the teams know that they're supposed to win (cover) at -6.5 the week before a bye?

You can back-fit just about anything to make it look good.

bombCanada · 10-30-09, 11:36 AM

Originally posted by TGoat

It's a back-fit, obviously. Seriously, do the teams know that they're supposed to win (cover) at -6.5 the week before a bye? You can back-fit just about anything to make it look good.

The original logic behind the trend was that favorites going into a bye weekend want to be able to look ahead instead of perseverating over a failure to cover for two whole weeks (14 days! 336 hours! 20160 minutes!) rather than the usual one week of self-flagellation after a loss.

Any really good team knows that they're supposed to cover 16 games a season.

TGoat · 10-30-09, 01:41 PM

Originally posted by bombCanada

The original logic behind the trend was that favorites going into a bye weekend want to be able to look ahead instead of perseverating over a failure to cover for two whole weeks (14 days! 336 hours! 20160 minutes!) rather than the usual one week of self-flagellation after a loss.

Any really good team knows that they're supposed to cover 16 games a season.

Sorry, but I don't follow. Are you saying that professional sporting teams think about covering the spread vis-a-vis winning?

And if we're talking football, I don't understand the reference to 16 games a season. Cover 16 games a season? Don't get it.

bombCanada · 10-31-09, 04:49 AM

Originally posted by TGoat

Sorry, but I don't follow. Are you saying that professional sporting teams think about covering the spread vis-a-vis winning? And if we're talking football, I don't understand the reference to 16 games a season. Cover 16 games a season? Don't get it.

I'm not saying anything. The theory says.

And yes, this is a pro football trend. So a 17 week season with 1 bye week would equal a 16 game season. It was sort of a joke.

And if you don't get it, don't bet it. I'm not.

It's a curiosity exercise.

jackpot269 · 10-31-09, 06:28 AM

Good idea to test drive a car before you buy it !!!!!!!!!!!!!!!!