[aljack]

I am going to keep this simple and am wondering what everyone here has to say about this...

-

SU-Trends are misleading. When team A is 25-6 in their last 31 games. A bunch of those games; team A is like a -200 to -250 fav. So those trends aren’t profitable - get a unit count on SU trends. Otherwise, put the work in and get your edges elsewhere.

The Detroit Tigers are a perfect example.

• Tigers are 24-56 in their last 80 overall.

So, most squares will look at this and think - "well if you Fade the Tigers you are getting a 70% winning rate."

So they'll take the Indians -210 vs the Tigers. They will lay 2u's and lose. Or they will take Indians -1.5 -120 and play the single unit - and lose - and wonder why... Well the reason is because tailing those trends are not +EV plays.

The proper way to tail a SU-Trend is by waiting for their to be value on that trend. Not on the RL - but on the ML. Taking a team on the RL or doubling up on a big fav based on these trends - you will lose more times than not.

If you think about it...

56-24 @ an avg line of -200 equals out to be 56 * 0.5 = 28 now minus the 24 loses.

So that gives you +4u's over pretty much half of an MLB season.

That is very close -EV.

Especially because most times - people are going to lay 2u's to win the 1u @ -200,

So now that equals.

(56 * 1 = 56) - (24*2 = 48) = +8u's

That is being generous @ -200 lines.

If you reduce that by even a slight margin, and add in the few times you are playing these games at 2u's @ -205, -210, -215, etc.

You will end up -EV, and -units

So please. When we're all on twitter, when we're on the youtube videos and on the forums. Can we please stop pretending to be sharp because we are copy and pasting trend lines from the various sites that provide them - because in reality. Those trends and those betting methodologies are very square and will end up with your bankroll being in the negative. I will promise you that.

-

Let me know what your thoughts are on this.

[Waterstpub87]

All trends are misleading. ATS ones are a waste of time too. I think anyone who studies this for more than 6 minutes knows that.

[aljack]

All trends are misleading. ATS ones are a waste of time too. I think anyone who studies this for more than 6 minutes knows that.

I agree they aren't predicative - but if you're gonna be lazy and just tail a trend - you may as well tail a trend that's +EV.

[Waterstpub87]

I agree they aren't predicative - but if you're gonna be lazy and just tail a trend - you may as well tail a trend that's +EV.

What trend would that be? Never heard of anyone having success with following any sort of trend.

[aljack]

What trend would that be? Never heard of anyone having success with following any sort of trend.

• Celtics are 47-22-1 ATS in their last 70 after scoring 100 points or more in their previous game.

That's: 47-22, +20.77u
If you just blindly tailed them after scoring 100 pts in their previous game you'd be up 20.77u's

That's +EV.

[Waterstpub87]

• Celtics are 47-22-1 ATS in their last 70 after scoring 100 points or more in their previous game.

That's: 47-22, +20.77u
If you just blindly tailed them after scoring 100 pts in their previous game you'd be up 20.77u's

That's +EV.

How did it do the 100 games before that? or if you took every team after then scored more than 100 points the game before? Or is this current version of the celtics just super good after they score an arbitrary 100+ points?

Do you really believe that the Celtics are undervalued so dramatically after winning?

Better explanation that this is just a random trend

[aljack]

I know it's a random trend - but if it's winning at more than 55% at -110, you are winning money off of it.

I track these trends and they become apart of my handicap - and if something is winning at a high rate and i am making money and consistently winning off of it - i am not going to stop betting it - but i am agreeing with you that more times than not you will not find a trend that is +EV.

I was simply just showing you the rare case when an ATS trend is +EV and worth tailing.

But i do agree with you most trends are absolute garbage - especially small sample size trends.

[KVB]

• Celtics are 47-22-1 ATS in their last 70 after scoring 100 points or more in their previous game.

That's: 47-22, +20.77u
If you just blindly tailed them after scoring 100 pts in their previous game you'd be up 20.77u's

That's +EV.

I disagree with this.

Your assessment on expected value, while technically passing the equation test, is ripe for flaws. The first and most obvious is your sample size.

Trends can be helpful if used correctly and because of the unsophisticated bettors out there, well publicized trends can and will skew the market.

Opportunities can be found there, but they are within a much more precise measurement of expected value.

[aljack]

I disagree with this.

Your assessment on expected value, while technically passing the equation test, is ripe for flaws. The first and most obvious is your sample size.

Trends can be helpful if used correctly and because of the unsophisticated bettors out there, well publicized trends can and will skew the market.

Opportunities can be found there, but they are within a much more precise measurement of expected value.

I see what you're saying.

Especially since knowing when to jump on such a trend before the rest of the market does is what makes a pro from a joe. I understand what you are saying about the unsophisticated bettors.

I can post all the trends I want - the fact that it's +money after 70 or so games doesn't really matter because until it reaches that high % ROI, it's a meaningless number.

[jbayko]

I agree they aren't predicative - but if you're gonna be lazy and just tail a trend - you may as well tail a trend that's +EV.

Huh? If a trend is +EV then it is predictive by definition.

[gdon44]

I know it's a random trend - but if it's winning at more than 55% at -110, you are winning money off of it.

I track these trends and they become apart of my handicap - and if something is winning at a high rate and i am making money and consistently winning off of it - i am not going to stop betting it - but i am agreeing with you that more times than not you will not find a trend that is +EV.

I was simply just showing you the rare case when an ATS trend is +EV and worth tailing.

But i do agree with you most trends are absolute garbage - especially small sample size trends.

You are only winning money on it if you bet it before it became a trend. That's why trends are dangerous. They don't mean a thing if you start betting after the trend is over....and you never know when the trend will end because it's not predictive.

[gojetsgomoxies]

I see what you're saying. Especially since knowing when to jump on such a trend before the rest of the market does is what makes a pro from a joe. I understand what you are saying about the unsophisticated bettors.

i finally posted about gonzaga being ATS-gold at home WCC conference games. like 25-5 or something. as soon as i posted, they went 0-3 right away.

another guy did the whole "cleveland is horrible ATS at home". i think they went 2-0 right after that.

i'm not saying these are "doom switch" things but in cleveland's case their losing ATS margin was consistently very small after being very large early in season.

on the boston "100 points" that strikes me as data mining and of course it would have taken you pretty long to catch on to that even if it was/is real. i think it's better to assess how boston does on ATS, SU and O/U winning losing streaks. that is much less prone to data mining although all that i've and others have mentioned is ultimating data mining.

[gojetsgomoxies]

i agree totally that SU streaks depend crucially on ML... isn't this how people saying they have 90% winning records? my pick for this next week is alabama to beat umass in football.

[Bsims]

on the boston "100 points" that strikes me as data mining

Calling this data mining is fairly accurate. Data mining is very dangerous when it comes to sports handicapping. For instance, I could write a program to pass past NBA data, looping through all teams, then through team's scores from 90 to 110, then see what they did ATS. It's almost guaranteed to find some combination of team, points, and number of games in a row that will show significant results ATS in their next game. And this is totally worthless.

The problem is that this approach is getting the "cart before the horse". You need to identify a theory, then use past data to test this theory to see if it holds up. Then be prepared to have this theory fail over 90% of the time.

[jbayko]

You are only winning money on it if you bet it before it became a trend. That's why trends are dangerous. They don't mean a thing if you start betting after the trend is over....and you never know when the trend will end because it's not predictive.

This is exactly right. People falling for trends is an extension of the "hot hand fallacy".

All bettors would do well to time to studying all of the various human cognitive biases that are working against them on every decision they make, most often without being aware of it. Each of us believes that we're always thinking rationally, but our brains have evolved to do some strange stuff that makes sense (or made sense at one point in time) from a survival of the fittest standpoint.

[u21c3f6]

The best advice I can give most is: Stop trying to pick the winner!

If you could select the winner every time, then certainly this advice is not for you. If you cannot select the winner every time, then regardless of your win %, the odds that you wager at will determine whether you are +EV or -EV.

Most of the "analysis" ( from datamining) posted is usually in the form of a high win rate. Of course as already posted by others, you don't know when to start or when to stop.

You would be better served if you analyzed the effect that such "trends" have on the odds to determine if there may be value on one side or the other. Yes, one side or the other. It is amazing to me that some cannot bet the other side regardless of the odds because they can't see their selected side losing.

Not to be a dick, but I love it when someone posts that a wager can't lose or is the bet of the year etc. when I am going to wager on the other side because I know I am on the "right" side and I will usually get an even better price because of other bettors with the same opinion. That doesn't mean that my wager is going to win. It just means that these wagers are +EV for me.

This is a much longer discussion but I will leave it here for now.

Joe.

[gdon44]

All bettors would do well to time to studying all of the various human cognitive biases that are working against them on every decision they make, most often without being aware of it. Each of us believes that we're always thinking rationally, but our brains have evolved to do some strange stuff that makes sense (or made sense at one point in time) from a survival of the fittest standpoint.

David McRaney wrote two great books on this topic. You are Not so Smart and You are now less dumb Both are excellent even for non-sports bettors. He also has a podcast I believe.

[gojetsgomoxies]

this is the kind of info/talk we should have more of on here... i.e. what is good actionable info?

[Bsims]

this is the kind of info/talk we should have more of on here... i.e. what is good actionable info?

Agree. This forum is the reason I gave up on the other sports wagering sites years ago and only visit SBR.

[Rich Boy]

Good post, I love this kind of stuff and have run backtesting systems and models for sports and stocks for years.

What Bsims said is dead on, if you look hard enough you will be able to find a certain set of filters that seems to produce very high win rates, but the samples are usually small. If you cant theoretically explain WHY these set of filters produce winners other than chance or luck then the system has no merit.

Also if a system has proven with backtests to be profitable, there is no guarantee that it will continue to be profitable, markets are always changing and adapting.

[TheMoneyShot]

The only true way to wager on sports and have success.... is give yourself some OUTS. Most don't have a plan after the first wager. You need to have a series of "WHAT IF'S" after your initial play is graded.

If you don't give yourself a series of outs... and a decent size bankroll... you won't win anything.

[Bsims]

The only true way to wager on sports and have success.... is give yourself some OUTS. Most don't have a plan after the first wager. You need to have a series of "WHAT IF'S" after your initial play is graded.

If you don't give yourself a series of outs... and a decent size bankroll... you won't win anything.

Did you mean to say after the "initial play is accepted"? The play being graded means the game is over. Then the outs would have to be on different games. I initially assumed you were implying additional wagers on the same game.

[Bsims]

Also if a system has proven with backtests to be profitable, there is no guarantee that it will continue to be profitable, markets are always changing and adapting.

For we back testers, this is the scary part. Recently I back tested another posters "Simple MLB System" for 8 seasons. It was a loser (as expected) overall and for 6 of the 8 years. The problem is that it was profitable in 2013 and 2014. If those years were used for back testing, one would falsely assume it was a good system.

Another example that is really troubling is the current system I'm using. This system hasn't had a losing year in the last 8 years with return per dollar being from $1.05 to $1.17. But, after 4 weeks of this season I'm down over $12,000. So do I continue staying with the system, or bail out?

[Rich Boy]

For we back testers, this is the scary part. Recently I back tested another posters "Simple MLB System" for 8 seasons. It was a loser (as expected) overall and for 6 of the 8 years. The problem is that it was profitable in 2013 and 2014. If those years were used for back testing, one would falsely assume it was a good system.

Another example that is really troubling is the current system I'm using. This system hasn't had a losing year in the last 8 years with return per dollar being from $1.05 to $1.17. But, after 4 weeks of this season I'm down over $12,000. So do I continue staying with the system, or bail out?

Sample size is everything and its difficult to differentiate between expected variance and actual expected win probability. It becomes even more difficult when you are emotionally and financially invested in a system.

Similar thing in the stock markets, you have a system and you put it into place. How much confidence do you have in such system? When do you trash it and revamp the whole thing? If you have early success how do you its not short term luck and variance? Ahh if we only had a crystal ball...

[TheMoneyShot]

Did you mean to say after the "initial play is accepted"? The play being graded means the game is over. Then the outs would have to be on different games. I initially assumed you were implying additional wagers on the same game.

Every sports handicapper should know their own strengths and weaknesses. What works for me... might not work for you.

I have a series of plays.... if my initial wager loses... and I have a series of plays after my initial wager wins. Typically.... the OUT games aren't on the same day.... but sometimes they are. Also, I may not even use my series of plays after my wager wins. I always like to recalculate and go with a FRESH BOARD for the next day.

Point being... majority of people who wager on sports... don't even have 4 games lined up after the initial wager. Your 4 games aren't chasers... they are calculated plays.... all different unit amounts.

[flyingcarbs]

For we back testers, this is the scary part. Recently I back tested another posters "Simple MLB System" for 8 seasons. It was a loser (as expected) overall and for 6 of the 8 years. The problem is that it was profitable in 2013 and 2014. If those years were used for back testing, one would falsely assume it was a good system.

Another example that is really troubling is the current system I'm using. This system hasn't had a losing year in the last 8 years with return per dollar being from $1.05 to $1.17. But, after 4 weeks of this season I'm down over $12,000. So do I continue staying with the system, or bail out?

Have you looked at what the maximum draw down was for previous years? Whats the current draw down, does it deviate from the average?
We know every trend reverses eventually, whether its unexpected and how temporary is the reversal are the important questions.

[Bsims]

Have you looked at what the maximum draw down was for previous years? Whats the current draw down, does it deviate from the average?

Excellent suggestion. I looked back at what would have occurred for each of the last 8 years if I had used this season's system. Following are the results.

Year	Lo Point	Hi Point	Final
2010	-$2,900	$34,110	$34,110
2011	-$1,350	$26,425	$14,835
2012	-$8,815	$9,530	$9,110
2013	-$605	$38,760	$25,445
2014	-$15,200	$25,485	$24,420
2015	-$7,905	$26,055	$14,970
2016	-$16,905	$13,580	$13,080
2017	-$6,460	$17,755	$12,285

This is a demonstration of the variability Rich Boy described. One year there was a low point of being down almost $17,000. On the other hand, there was a peak over $38,000 in one season. It's time to keep the faith and ride it out.

[El Terrible]

David McRaney wrote two great books on this topic. You are Not so Smart and You are now less dumb Both are excellent even for non-sports bettors. He also has a podcast I believe.

My handicapping is like my golf swing, i'm not messing with it...and Ive never considered trends.

[john hardesty]

I concur, remember trends bend one way for no more than 7 games, then they bounce the other way!

[john hardesty]

Sir, remember, Vegas takes it all: trends, systems, and hunches!

[El Terrible]

"Vegas" (whoever that is?) doesn't know who's gonna cover ATS, they pay the winers from the losers and take a cut for holding everyone's money...

[HeeeHAWWWW]

For we back testers, this is the scary part. Recently I back tested another posters "Simple MLB System" for 8 seasons. It was a loser (as expected) overall and for 6 of the 8 years. The problem is that it was profitable in 2013 and 2014. If those years were used for back testing, one would falsely assume it was a good system.

Ha, that's so familiar to me.

It's why backtesting isn't just about reaching an adequate sample size - sometimes you get these market/result structural alignments, where your model's biases match those of a particular season (eg simplistically, you produce lots of overs, and it happens to be a single season with ..... lots of overs).

[Bsims]

Ha, that's so familiar to me.

It's why backtesting isn't just about reaching an adequate sample size - sometimes you get these market/result structural alignments, where your model's biases match those of a particular season (eg simplistically, you produce lots of overs, and it happens to be a single season with ..... lots of overs).

It's particularly concerning if you've used past data to find something interesting to examine. Then you use that same data to back test. One of the pitfalls of data mining.

One technique I've used is to segment past data. For instance, build two MLB data files, one with games played on days with even day numbers and the other for odd number day numbers. Data mine one to find patterns, then the other to back test.

[El Terrible]

I don't get the need to even look at systems. If you don't want to look over games and come to your own conclusions on them, why bother?

[Bsims]

I don't get the need to even look at systems. If you don't want to look over games and come to your own conclusions on them, why bother?

There are two reasons. First, you can cover all the games. Looking at each game is time consuming and your available time might mean you only look at a few. Second, some of us view this as an investment. As such we prefer hard analysis over gut feels.