Runline plays

So I've been checking out the stats on runlines this year and found that the teams that lead the league in run differential are also winning between 40% and 50% of their games by 2 or more runs. Also, the teams that have the worst run differentials are losing 40% to 50% of their games by 2 or more runs. With the avg runlines well above +120, these would have been cashing.

So what if you played each of these teams on the RL all year long?  I back-tested this idea against last year's stats for the best in run diff and the worst. Here's what I found. The best teams won 76 games (nyy), 75 games (phi) and 77 games (tex) games each. This means, for example, last year the yankees were 76-86 in these plays with a break-even mark of about +114 on average. On the other side, the teams with the worst run differential spread last year also lost a lot of games by 2 runs. The 3 worst lost 67 games (min), 78 games (hou) and 71 games (bal). Fading the min RL you would have gone 67-95 needing a break-even mark on average of only about +122.

The teams leading in run differential so far this year are tex, stl and atl while the teams with the worst run differential right now min, sd, mil and kc. I think if you play these on the RL over the rest of the year you could come out ahead, especially if you keep updating your list of teams so you are always following the best and fading the worst.

Today, the plays would be
tex -1.5 +105
stl -1.5 +140
atl -1.5 +155
bal -1.5 +195

and to fade the worst, you would play
tex -1.5 +105 (fade min)
was -1.5 +150 (fade sd)
nym -1.5 +245 (fade mil)
cle -1.5 +130 (fade kc)

I'm going to track these plays for a little while and see if these trends keep up.

actually, the cle play fades min and the tex play fades kc.

I like the idea

if you played each of those teams all year round, as your commentary indicated, on the RL, based solely on what you have brought forward, without description of a handicap; you would delete your bankroll before the playoffs started, interesting assumption, games are won by handicapping them individually..

may you prosper always-

Based on the numbers I brought forward, I beg to differ. Please elaborate on your argument. Thanks.

I had a chance to do some more backtesting and was able to check 2010. The results for that year are consistent with those of 2011. The top 3 teams that led in run differential also won over 40% of their games by at least 2 runs. They were tam (68 wins), nyy (74 wins) and phi (69 wins). The bottom 3 teams in run differential at the end of the 2010 season were pit (lost 74 games by 2 runs or more), hou (lost 68), and laa (lost 56). These all meet a 40% win percentage except for the angels (35%). If the angels had been included one would have needed to average a line of about +145 to break even. The next lowest teams (hou or tam with 68 wins each) would have needed an average line of about +120 to break even.

Going off of these numbers for 2010, 2011 and the first part of 2012 I think there should be a cut off RL winning percentage of 40% and above. If a team falls into this category, they will be a play for that day.

Keep in mind, as I think sports-Network was trying to point out, that this system is not based on individual daily matchups but RL win percentages and current records only. The real challenge of this system is to pick the teams that will continue to win by 2 at a 40% clip or higher.


thanks

Jonny, I think the problem is you don't know the teams future records. Of course your back testing working because top teams had the most run differential. Just my 2 cents but GL. 

Are you backtesting the teams that finished the year with the highest and lowest run differential? If so you're backtesting isn't necessarily reflective of what teams had the highest and lowest run differentials in May when you're betting them. Just throwin' it out there, GL to you.

You're absolutely right, hyvong. The tricky part is to predict which teams are going to stay at the top and bottom of the run differential stats. Right now, there's no way to tell who's going to finish 1, 2 ,3, 28, 29 and 30th in that stat. Nor do we know how the teams we've chosen are going to continue to perform. That's something I've been wanting to look at further. So I took a closer look at the past year's run diff stats. That leads me to the significant point BCap88 brought up, which was much like yours I think...

... You are totally right, BCap88. That is exactly it. The teams change throughout the year. But by how much? And how often? That's what I wanted to find out. So I tracked the top 3 teams and the bottom 3 teams from last year, on a week by week basis. Here's the data . Basically nyy, bos and phi were the top 3 for most of the way with cle, stl and atl sneaking in and out with tex only joining in the last 2 weeks for the most part. Those top 3 combined for 371 games, going 164-207 for a 44% winning percentage and a break-even percentage of a +126 line average. Now keep in mind, this is by adjusting the teams once a week, not daily. As for the bottom 3, they pretty much followed the same pattern with hou, bal and min being the teams most of the way. The bottom 3 combined for the 372 games, going 149-223 for a 40% winning percentage and a break-even percentage of a +150 line average. A little worse I would admit. Altogether, the top 3 and bottom 3 combined for 743 games, going 313-430 for a 42% winning percentage and a break-even percentage of a +137 line average.

I guess the next question is did those wins average above a +137 line. That's prolly another backtest for another day. But I will say that after tracking those teams I feel pretty confident about being able to pick the teams that will give us a 40% and above return, consistently for the rest of the year. They don't rotate much.

Thanks for the questions and let me know if you guys think of something else. And don't forget to check out the data. There's some interesting trends in there.

Yesterday, the mets and the nats came through with the mets paying off big. I'll start tracking these out on a spreadsheet later but for now just know that the system went 2-6 yesterday for a loss of about .05 units.


To play the top 3 today, take

tex -1.5 +135
stl -1.5 +125
lad -1.5 +195


and to fade the worst, play
cle -1.5 +120 (fade min)
was-1.5 +130 (fade sd)
nym -1.5 +270 (fade mil)



gl

whoa, my bad. Tex is actually a -120 for a -1.5. (I played 2 units because of low juice)

Johnny;

Great stuff.  Have you thought about a filter to exclude games, for example, with a low total like 6.5 or 7?  My experience (and I can't backtest this) has been giving up 1.5 runs on the RL when the total is low makes it a little harder to cover.  Today's Dodger game is an example.  The total is 6, so this game might be filtered out.

great idea, Budman717. I've been trying to think of variables of this system that could be filtered. That could def be one. It would be a bitch to backtest and check all those games, tho. But if I could, I would find out (A) what the actual lines were for each win and (B) what the numbers would be like if games with an O/U of 7 were eliminated. I do suspect that the lines for each of those games were reflective of the low totals so you would also eliminate some of the big wins. The dodgers, for example, were +195 when I booked the play earlier.

Do you think it would increase your win % if you filtered the plays on the highest run differentials to play on them only if they were playing teams with negative RDs? And vice versa for the fades? Playing on teams with positive RDs against the bottom 3 teams?

Nice suggestion, BCap. I was starting to cool on this system but these filter ideas have got me a little intrigued again. I think between everyone's ideas we may have a viable system here. So far, here's what it's shaping up to be:

To be a RL play -

A) Must be top 3 team or bottom 3 in run diff for that day.
B) Must have O/U of 7 and above.
C) Matchup must be positive run diff vs negative run diff.

Also, I was thinking it might be helpful to consider the ERA matchups and power ranking matchups of each game. Let me know what you guys think or if anything else comes to mind. I'll start official plays once these filters are all set.

thx to all.

gl

Sorry to be a pain, but I see that the last three days if you were playing the RL+1,5 line for all dog team with odds>-130 you  would be 10-2 
Last night all 4 teams with the better odds at dog+1,5 line won!  That is SD, HOU, BAL and KC.   Heavy favs are not a good choice lately.   Obviously, setting filters is good because you avoid phillies (they are the biggest fav with the coolest bats).  keep it up, soon enough this will change!

You're not a pain, bro! Whatever comments you have are totally welcome here. I noticed some bad streaks when i went thru the data from 2011 so I knew there would be some nights like that. And after another closer look this morning, it looks like a this system needs more than a few filters. I ran the numbers on the nyy for last year and they came out ahead only about 5 units after all of those plays.

On a side note I did notice how well teams bounce back from a shutout, winning 1 it's next 2 games. It's like 75%. Crazy, esp since about half of those lines are dog lines.

Oh yeah, I forgot to mention... nice angle. I had been wondering about the +1.5 side. And, yes, those philly bats are dead this year!