I know it is baseball season but...

I was commenting on a topic in another thread and actually saw something that I really hadn't noticed before, even though I have argued the topic endlessly, and from many many angles.

But in terms of the -3 line and its closing value and results, and the place I keep track of anyways.

I have data going back to the 1985-86 season. So roughly 20 years. I have broken this -3 topic down on a league wide basis, as well as a team by team basis. (finding that about 7 teams are involved in the majority of pushed). But what I didn't really value was how many more games there are lined -3 now than there used to be. I knew it was more, but didn't really take stock in it until I was looking this morning. So I broke it down.

Friom 85-86 to 93-94 season, there were 221 games closing -3. Of those, 20 laned on the 3. So right around that preconceived 10% mark most like to quote.

But from 95-96 to last year, there were 388 games closing at -3, of those 45 landed on the 3. So a quick math test shows that is about 11.5%.

Now I have come on here time and time again saying that -3 is vastly overrated, and it is in my opinion still, despite having the past 10 years show a "rise" in "probability" from 10% to 11.5% , as well as more games being lined at that number (roughly an 80% increase in total games) SO logic would say that books are right and the 3 is worth all that extra.

I say bull shiit. I say the books get a break because 3 years wre way out of the ordinary. From 86-94 it was like clock work, 1-2 games a year out of 20-25. One year there wasn't a single push on 3 but there were also maybe only 17 games that year.

But since 94 the results are crazy. One year might have 3, another might have 7, another still might have more. Well there were 3 yerars in particular that blow it out of all proportion, and basically make it impossible to put a real value on the number, unless you call tjose three years anomolies, and the other 6 years the norm", which is probably a stretch too.

But in 97-98 season there were 32 games closing -3, of those 7 landed on the number. In 99-00 season there were 31 games that closed -3, of those 8 landed on the number, and the grandaddy of them all was 01-02 season (the years Aces went under) where 11 of 41 games landed on the 3 (one site I check has the push number at 14 of 47, and another has it at 13 of 50 so a wide range of results for that year, but regardless it was way out of normal).

Now the discalimer, I made a math error in the other post it wasn't 24 of 80 games, it is 26 of 104. My programbreaks it down in away/home, and one year where were 24 road games unaccounted for, in those 24 road games there were 2 pushes, ths the discrepency.

But regardless of that, those 3 years are enough to push the rate of expectation from 6.7% all the way to 11.5%. In those 3 years you have 26 games pushing(from my source, more or less than that at others I am sure) But in the other six years there were a combined 19 pushes. Also in those six years you have 284 games that closed at -3.

So what is the norm? And can it be determined? The three years that seem to be anomolies had an average of 35 games closing -3, the other six years had an average of 47. Does that make a difference? It certainly does when you start averaging results out, the more total games compared to fewer target results drastically changes the percentages.

So one year you might have 5% of game landing on a push, another you might have 25%. With such a huge discrepency canone even determine if a "fair" expectation can be given?

Yes, prior to 94 I think it could have. But since then, I do not not think so. SO assuming anything is bad in this case.

It is sort of like running across traffic. "statically" you havea 10% chance to get hit by a car. But at rush hour that raises to probably 50%, but at midnight it is probably 2%. But unfortunately with this subject it isn't as simple as timing issue, it is just whether or not things happen in the right sequence.

But I can say that trying to vaue a game on -3 is not as simple as saying it is worth this or worth that, because the results are way too volatile. Now I am sure others will come on and say that LONG TERM results are what gives you a probability/value factor. I agree, but not when there has seemingly been no "pattern" to those results for at least 10 years, and no way to say what is normal and what isn't.

Thanks for sharing that info Sir.

Could it possibibly be something to do with the way the NFL has evolved over the last 6-7 years, certainly games are a tighter. Although I have accepted a value of 10.9% for the 3, I have always thought that figure over-rated from a betting point of view. I have great difficulty laying a 3 at that worth on the M/L.

Parity has changed alot in the NFL. Hmmm, maybe not. Look at the NFC. :lolBIG:

You can't see the forest for the trees want, you never will.

Never has someone confused such a simple issue.

Pancho, you actuially touch on a good point, and one that a lot of people do not understand...


When is trying to make something more difficult than it really is right/wrong?

For example. Say the league average is in fact 10%. But what if there are teams out there with not a single push on a -3 number eiother as a favorite or as a dog?

Is that making it more difficult or less difficult? Assuming that 10% of their games will be pushes, and they have played say 40 games lined at 3, and not pushed once, with no pushes. Does that mean they are due for a run on pushes? Or is it something else?

When "handicapping" or in the very least predictingh a likelihood of a push,. or a value of a certain number I think having Spcific teams results IS in fact a MAJOR part of detemining value.

If a team has NEVER pushed on a game lined 3, that to me means a lot. If a couple teams seemingly push every time they are in a game lined 3 THAT is also a MAJOR factor.

If The Raiders have pushed in 20% of their games, and The Bucs have pushed in 0% of their games, the AVERAGE is still 10%, but is it really indicitive on each teams "probability" of pushing?

It goes deeper than that, it actually could be extrapolated to all lines and all teams, that is obviously quite an undertaking, but programs do exist that show how teams fare at certain point spreads. I can break down line by line data for every team in the league, but since I don't bet a lot of one way games it doesn't do me much good, and not sure if it would anyways. But I am sure it would be useful for some to know how teams do in certain perameters. As in from a -1 to -3 fave compared to a 3.5 to 6.5 fave, and so on.

But is it more confusing or is it more extensive? As in situational. Baseball is built on sitations...hit and run, when to bunt, when to pitch a lefty or a righty, when t pull a guy on a roll in favor of a pitching match up that is favorable. betting should be no different in my opinion. Especially when it comes to trying to determine "value". Since Value is relative, having some set outlines definately helps solidify what kind of value if any one team has over another.

What about 2-point conversion ? I'm not sure what year that came in to play, but could that account for part of the difference ?

I'm amazed at Wanit's posts

He has 20 years of data. It convincingly shows the incidence of NFL games lined at 3 landing 3 is 10-11.5%.

Obviously, games lined with low totals "hit" slightly more often, those with high totals slightly less often. But all things being equal, 11% is a damn good estimate of how often we expect an NFL -3 chalk to win by exactly 3.

Maybe it's 10.8%, maybe 11.7%. But anyone arguing the number on any individual game shoud be anywhere other than 11% isn't going to get picked up on any sharp betting team's waiver wire any time soon*.

*Standard caveat, there are always the very rare exceptions to any rule. Example would be a game played in a hurricane where it was impossible to kick field goals, etc. Also, obviously low total games closer to 12%, high total games closer to 10%.

Fezzik


PRE 94 it is like clockwork, between 20-30 games closing at 3, and between 1-4 games landing on the 3. With so few games landing on 3, then one or two here or there extra doesn't seem that big a deal. But even when more hit, it WAS proportionalm, I will give that.

But POST 94 the numbers are all over the board. With no rhyme or reason. I will say that PRE 94, YES, the "prtobability" of having a game land on 3 that was closed at 3 was 10%. Not only do the number back it up but it was CONSISTANT.

Obviously in sports gambling, consistancy is a very relative and debateable point. But when it is/was as "locked in" as I found, it is definately consistant.

But now you have one year that might be 5%, another that might be 4%, another that is 25%, another that is 16%, one is 7%, where is the consistancy there? There is none. The only cosistant factor I found were in the years where there wasn't an "anomoly", the years where it was between 5-7%, granted there are only six of those years, but the other 3 have been a lot higher.

In other words, the ONLY reason it has maintained a 10-11% AVERAGE/MEDIAN is because the results are so extreme. It is one thing to see 9%, 11%, 10.5%, 8.8% on a "consistant" basis, and quite another to see a hodge podge of results that ONLY WHEN AVERAGED equals 10-11%.

IMO recognizing THAT helps someone determine a more "true value" that just taking a radom average and go from there.

It is akin to saying that nearly half the people on the planet are Indian or Chinses, and then trying to assume/say that half the people EVERYWHERE are Indian or Chinese, it simply does not correlate out that way. There are a few intermixed, but to assume that if you take 10 people from around the globe and in them you will find a mix of 5 Chinese and Indians is flawed, just like it is flawed to assume that if you take a sample of 20 games closing at -3, you will find 2 that pushed on that number.

That is all I am getting at here. PRE 94 one might be hard pressed to argue, but post 94 it is a simple pint to make.

As far as totals, I have not fouind any correlation between posted totals being higher or lower having any effect on "push probability", before or after the fact.

Go flip 40 coins and chart the heads/tails

Do this for 20 trials. Report back on the "patterns".

You will be amazed at the runs you will see. 5 years of "expected" data, then two years of 25 and 27 heads, etc.

In small date samples, random variance will always be all over the planet.

Coin flipping is truly random.


A -3 line in football is not. first one has to assume the book has put up a "legit" number i.e. one that is more predicitive than one that just a number to balance action, or one to just be able to overcharge for one side over the other if they know that one side will take more action.

When flipping a coin, the only real variable is the speed or pressure with which it is flipped, (and maybe wind) but even so there are basically only 2 options a head or a tails.

In sports there are an infinite number of possibilities.

I understand this isn't really what you were getting at, but it does come into play.

But trying to compare a year in football with 15 games that push and then looking at the next year that has maybe 2 in the same amount of (flips) games ins't a good example IMO.

Maybe trying to guess heads or tails and tracking when you are right and when you are wrong might be better, but even then, there was an outside infklunce initially (the linemaker/book or move that makes the line close at 3 in the first place)

I have said it before, that if books lined EVERY game -3 They could actually make more money in the long run. They simply do no need to do so. Obviously then they would get a lot of pushes, but when they started charging -150 for a side that most might agree should be -7, (but wouldn't bet the -150 number) they would make it up on the other side. Espcially when you look at see those teams that are lined -6/6.5/7 losing SU or winning just by 3. So the vig would make it up for them with the losers, and the pushes would protect them from the guy hammering sides that are seeming "steakls" at only -3.

It is akin to the big faves in bases getting taken -1.5 on the RL. A lot of guys g do it, but even when the book isn't "protected" with a push they (the books) still make money doing it. If they were geting killed they would manipulate them more than they do.

Lines are random. Any line made. Might have reasoning,might have variable input as to why it is what it is. But in the end it is still random. In pro sports where there is a pointspread system, the points only matter about 17-20% of the time. Meaning that if the dog doesn't win SU, then they only cover about 9-10% of the time, and get a push between 8-10% of the time.

College is slightly better for dogs, but not much. And really only because getting pushes in the college games is far less frequent than it is in the pros.

This can be debated from all angles, and some will be right in some points, some will be right in other points. But ultimately, in my opinion one cannot value a -3 line anymore than they can value any other.