"Past Performance is Not Indicative of Future Results"

[JohnAnthony]

This is confusing me more than a bit.

What is the value of stats if this statement is true? should we combine them with trends, inuries, etc...to come up with a pick?

And what is this "past"? if I'm only looking at Team A's Average Goals Per Game for this season, how much of a "past" is that in regards to the statement?

*sits and waits for FourLenghtsClear and mathdotcom*

[GarbageMan]

Without past performance, push charts wouldn't be applicable, so yeah I see your point.
*not waiting for matho to respond*

[mathdotcom]

Past performance is of course indicative of future results.

Why does everyone think the Pats or Packers or Saints will win the superbowl in the future? Because they kicked the shit out of everyone in the past.

[PLF]

Well the truth is, past performance itself isn't indicative of future results BUT we all also know that 'History tends to repeat itself'. So that's where stats and past meetings, form, past results, etc. come in.

So both sayings have their merits I think and to gauge and judge which applies more to each given situation is probably the fine line between being profitable in the long run or not.

Just my 2 cents.

[JohnAnthony]

Thanks folks. Of course I'll appreciate everyone's input, I just mentioned these two guys because they've been very helpful

[YouMama]

if u believe "Past Performance is Not Indicative of Future Results" ... how do u come up with any factors for any kind of anylasis?

[FourLengthsClear]

It's a saying that might be true when talking about mutual funds but does not apply to sports betting.

If team A beat team B 4-0 last week, does it mean that they will win 4-0 if they are playing again tomorrow? Of course not bit is it relevant information? Yes.

If that 4-0 was not last week but five months ago (still same season), is it as relevant?

In soccer you often hear stats like Team B has not beaten Team A away from home for the last 25 years. Is that something that we should be taken into account or perhaps more importantly might it be reflected in the odds that are on offer?

The answers to these questions vary a bit from sport to sport and, of course, are dependant on situational aspects but how you go about answering (and quantifying) them will be important going forward.

[JohnAnthony]

Guys, my beliefs concur with yours strongly. But I remember reading one authoritive member (not sure exactly who now) who said the exact same thing. Again, I've been using and modifying my stats with GREAT success (a bit less than 500 picks, small sample size), but for the last week it was kinda stuck in the back of my mind.

Thank you all!!

[GarbageMan]

Stats like man utd are 7-1 ats when playing on a wed under floodlights and Wayne has banged coleen the night before are obv useless. When you get a whole database of how every team in a league had performed against a certain spread over a large sample size, well then you have some useful info

[jerry]

Stats like man utd are 7-1 ats when playing on a wed under floodlights and Wayne has banged coleen the night before are obv useless. When you get a whole database of how every team in a league had performed against a certain spread over a large sample size, well then you have some useful info

If Wayne banged Colleen the night before, and if the betting public see's this as important info, it will be reflected in the odds. Therefore, anything that happened in the past is important.

[YouMama]

... look what happened w Matt Kemp when he started banging Rihanna

[Salmon Steak]

What will impact the results of this game the most? Current situational information? stats from the past? both? Other? Should we weight something based upon its purported impact? I know what my answer is, but it might be different for someone else. Go with your gut friend.

[TheLock]

Past performance sets the line

Models and relative value predict future results

[hutennis]

Past performance is of course indicative of future results.

Why does everyone think the Pats or Packers or Saints will win the superbowl in the future? Because they kicked the shit out of everyone in the past.

Sure, past performance is indicative of future results in sports. Favorite in the past is most likely the favorite in the future. The only problem is that this knowledge does not do us any good as far as making money.

Sure, if only we'd get paid 31 to 1 for picking Superbowl winner, we would all be rich betting on Pats, Packers or Saints. But we are not getting that much, are not we?

Every little bit of value we may find in knowing past is being promptly sucked out by assigned implied probabilities of the result and farther sharpening of those probabilities by open market.
A glimmer of hope that some minor "inefficiencies" may still remain is most likely being more than negated by yet additional premium applied - juice.

It's simply amazing how many otherwise smart people have problem grasping such an exceptionally simple concept.

All that being said, I have to report, that after looking at more than 30 years worth of data across all major sports both in sides and totals (millions data points) I think I have found a remarkable indicator/pattern. I did a regression analysis too and it seems to have a perfect correlation. I found that team that is being posted in odds table (american format) with higher absolute number wins (covers) by far more often. As far as I see it, this is a mathematical fact. If you don't believe me, check it out for yourself.
Now, with that knowledge in our pockets, let's go ahead and make some serious dough!!!

[CHUBNUT]

Stats like man utd are 7-1 ats when playing on a wed under floodlights and Wayne has banged coleen the night before are obv useless. When you get a whole database of how every team in a league had performed against a certain spread over a large sample size, well then you have some useful info

Where can you get the banging stats, they could be useful.

[parlayin]

I found that team that is being posted in odds table (american format) with higher absolute number wins (covers) by far more often.

Can you explain this? The team with more wins covers the line more often regardless of spread? So usually the favorite will have a better record (more wins). Are you saying that blindly betting them will be profitable?

[mathdotcom]

There's no shortcut. You need to decide what part of past performance is relevant. Halladay pitching really well on odd numbered days of the month is irrelevant, and team vs team records especially in a sport like baseball are incredibly overrated.

[hutennis]

There's no shortcut. You need to decide what part of past performance is relevant.

Although I do agree with this statement in general, I guess, I have to ask you to define "relevant".

To me "relevant" would be part of past performance that became unnoticed for the rest of the world and thus was not priced in already, making number I see in odds table incorrect and exploitable.

To determine that I need to know composition of factors that did went into establishing the number i think is exploitable.
Because it very well may be that something i consider overlooked is in fact there plus something relevant that was overlooked by me is there too, thus explaining the difference between two lines and in fact making my line inferior.

Checking against the "less assumptions - right hypothesis" rule it easy to see that assigning mistake to the world is not a rational choice to make.

How do you deal with this argument?

[mathdotcom]

Determining what's relevant is what separates the squares from the sharps. But the more you know about stats and the sport, the more successful you will be.

Love watching the squares in Vegas saying shit like 'the last time the Colts won in Chicago was 1978' and thinking that means something.

[That Foreign Guy]

It's a legal disclaimer to reflect the fact most mutual fund managers are flipping coins and the hard work is done by the marketing departments. They sell you with 17% returns in the past and cover their ass by saying it's not a guarantee and in fact may not even mean anything.

Past performance IS indicative of future results.

[Rufus]

It's a legal disclaimer to reflect the fact most mutual fund managers are flipping coins and the hard work is done by the marketing departments. They sell you with 17% returns in the past and cover their ass by saying it's not a guarantee and in fact may not even mean anything.

Past performance IS indicative of future results.

Actually, most mutual fund returns fail to keep pace with the market BEFORE fees. Meaning mutual funds are flipping an unfair (to themselves) coin.

Big picture, you can find lots of crazy-ass trends that are statistically significant but have no predictive value. You need to have some intuition for why you're seeing what you're seeing. Strictly data mining will help you build great explanatory models...but awful predictive ones.

[hutennis]

Strictly data mining will help you build great explanatory models...but awful predictive ones.

Exactly. People simply falling for a hindsight bias hard.

[Inkwell77]

there are past situations that are indicative of future situations

being able to decipher and put some sort of value on that edge is what situational sharps do (I assume)

everybody already knows this

[EVPlus]

In sports wagering, past performance can be indicative of future results (dependent on the variables applied).

However, past performance does NOT guarantee future results.

This isn't just semantics. This is an important distinction.

[Sportsguy_USA]

In sports wagering, past performance can be indicative of future results (dependent on the variables applied).

However, past performance does NOT guarantee future results.

This isn't just semantics. This is an important distinction.

Yep - that pretty much sums it up. Apply the same statement to non betting scenarios...like your investment / RRSP / 401K manager. I put my money with guys who have shown strong past performance, using statistics, trends, current events, etc. While not guaranteed of future earnings...my risk tolerance is lessened and my probability is higher to make some $. I apply the same to sports - whether on specific teams, players or cappers.

[High3rEl3m3nt]

Using baseball as an example, stats are nothing more than a range of expectation. They obviously don't predict the future, but they will likely provide two plots, where the actual number will fall between. In other sports, it's much harder to analyze stats to predict a likely outcome...just my two-cents.

[High3rEl3m3nt]

I also think baseball's application of stats is more indicative of the future because baseball is a series of individual actions. At any one point, only a few players are involved in a single play.

[That Foreign Guy]

Actually, most mutual fund returns fail to keep pace with the market BEFORE fees. Meaning mutual funds are flipping an unfair (to themselves) coin.

Yeah, friction / wanting to look smart / risk aversion is a bitch.

[mikew]

Big picture, you can find lots of crazy-ass trends that are statistically significant but have no predictive value. You need to have some intuition for why you're seeing what you're seeing. Strictly data mining will help you build great explanatory models...but awful predictive ones.

very clear, thanks