MLB UNDER System - Solid SDQL Results

[nfl_huskers]

Back in early April, I was looking at MLB Totals, and thought about the following situation that may generate more bets on the OVER, and possibly give more value to the UNDER:

Game Total is 8.
Pitcher 1 has a season-to-date ERA of 4.25.
Pitcher 2 has a season-to-date ERA of 4.50.

I wondered if bettors would flock to the OVER simply because 4.25 + 4.50 > 8.
I know most pitchers don't pitch an entire game, and this doesn't account for the bullpen. Again, just an idea.

So, I used totals and pitcher ERA to research via SDQL. The following simple query produced an immediate 5% ROI when betting the UNDER.

H and 6.5<=total<=8 and STDSERA>=total/2 and o:STDSERA>=total/2

The sweet spot on Total was from 6.5 to 8.0.
Both pitchers had an ERA of at least half the total.
The 'H' is just to pick a side; otherwise, games are returned twice.

UNDER Results since 2007:
Note: April is usually the slowest month, but this year it's on fire.

Games W-L-P (%under) Avg Total $ Under Season
20 11-9-0 (55.0%) 7.7 +$135 2007
74 42-26-6 (61.8%) 7.7 +$1,365 2008
68 35-31-2 (53.0%) 7.7 +$60 2009
99 54-42-3 (56.2%) 7.4 +$780 2010
233 123-99-11 (55.4%) 7.4 +$1,500 2011
249 134-100-15 (57.3%) 7.5 +$2,410 2012
233 119-100-14 (54.3%) 7.5 +$975 2013
269 143-114-12 (55.6%) 7.3 +$1,850 2014
330 172-145-13 (54.3%) 7.4 +$1,444 2015
187 103-77-7 (57.2%) 7.6 +$1,789 2016
30 22-8-0 (73.3%) 7.6 +$1,305 2017

Here's the monthly breakdown:
Note: Before this year, April was at -$269. Solid results the entire season.

Games W-L-P (%under) Avg Total $ Under Month
262 138-116-8 (54.3%) 7.5 +$1,036 4
324 165-143-16 (53.6%) 7.5 +$886 5
291 167-115-9 (59.2%) 7.5 +$4,158 6
260 143-104-13 (57.9%) 7.5 +$2,811 7
304 156-130-18 (54.5%) 7.5 +$1,427 8
313 169-125-19 (57.5%) 7.5 +$3,270 9
38 20-18-0 (52.6%) 7.3 +$25 10

Based on the hot start to the 2017 season, I figured I'd put the query out here, so it could be tracked the rest of the season. This will most likely jinx it.

There's actually another play tonight (5/1/2017): Tampa Bay at Miami with Total = 8



For those looking for fewer plays and an ROI better than just 5% ...

After some additional filtering on the ERA, and adding an extra filter for WHIP, I have the following system for betting the UNDER:

Bet the UNDER if ...
Total is 6.5 to 8.0
Both pitchers have an ERA of at least half the Total, but no higher than 5.25
Both pitchers have a WHIP from 1.25 to 1.55

H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2007


Results from last 10 seasons:
ROI: 13.0%
PROFIT: 72 units


Games W-L-P (%under) Avg Total $ Under Season
5 3-2-0 (60.0%) 7.7 +$100 2007
29 16-11-2 (59.3%) 7.7 +$380 2008
28 13-14-1 (48.1%) 7.7 -$215 2009
33 21-11-1 (65.6%) 7.4 +$905 2010
79 41-34-4 (54.7%) 7.4 +$375 2011
72 42-25-5 (62.7%) 7.5 +$1,455 2012
44 25-16-3 (61.0%) 7.6 +$815 2013
90 52-35-3 (59.8%) 7.4 +$1,315 2014
84 48-30-6 (61.5%) 7.4 +$1,500 2015
41 24-16-1 (60.0%) 7.7 +$632 2016

First game meeting the query requirements is typically the last week in April or first week in May, and it looks like the first potential plays are showing up for tomorrow (5/2/2017) right on schedule.

20170502 Cardinals Brewers Total=8.0
20170502 Padres Rockies Total=7.5
20170502 Tigers Indians Total=8.0

I'll follow up tomorrow if these become official UNDER plays.


For now, let's see if tonight's UNDER play on Tampa Bay at Miami with Total = 8 actually pans out.

Looks like I'll also need to remember to use tables when posting data, but no time to fix that right now.

[barryt]

Interesting results.
As you pointed out you were looking for an OVER result but the answer was an UNDER. Can you venture a guess as to why?
I'm no baseball expert( browsing here looking for something to bet on) but perhaps books have a algorithm that uses the same info but then sets the line too high for those pitchers, but low enough to get bettors to favour the OVER. In that case why would it only work for low ,8 or less, totals?

[nfl_huskers]

Interesting results.
As you pointed out you were looking for an OVER result but the answer was an UNDER. Can you venture a guess as to why?
I'm no baseball expert( browsing here looking for something to bet on) but perhaps books have a algorithm that uses the same info but then sets the line too high for those pitchers, but low enough to get bettors to favour the OVER. In that case why would it only work for low ,8 or less, totals?

Still researching. Hard to tell at this point.

1. Not enough emphasis on the bullpens? If the starters pitch a 'typical' game in-line with their ERA, but only pitch 6 innings, and their bullpens have a lower ERA, then the actual result could lean towards the under. Some initial filtering shows the ROI of the system goes up if the season-to-date ERA of the bullpens is lower than that of the starters.

2. Not enough emphasis on the opposing team's current offense? Maybe the pitchers are facing good teams in a hitting slump, but the Total is set based on their higher overall season average.

In general, those 6.5 to 8 totals just happen to be the sweet spot for the UNDER since 2007. For example, just betting all games with a Total of 6.5 since 2007 has an ROI of 5.9% ... total=6.5 and H and season>=2007

[Scrivero]

TB-Miami went under. Welcome huskers, very nice to see you here! There are some brilliant SDQL and MLB experts on these boards.

[TechnicalTrader]

Good stuff. I watched your thread across the street but was a bit too busy here to follow or provide feedback. Good stuff and thanks for bringing it over here!

I am guessing the reason why the unders are generating a nice RoI is simply because the market is undervaluing these pitchers. I'll see what I can find.

[TechnicalTrader]

BTW, I am getting timed out with this query. Is it correct or is sportsdatabase just being slow??

H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2007

[2daBank]

Still researching. Hard to tell at this point.

1. Not enough emphasis on the bullpens? If the starters pitch a 'typical' game in-line with their ERA, but only pitch 6 innings, and their bullpens have a lower ERA, then the actual result could lean towards the under. Some initial filtering shows the ROI of the system goes up if the season-to-date ERA of the bullpens is lower than that of the starters.

2. Not enough emphasis on the opposing team's current offense? Maybe the pitchers are facing good teams in a hitting slump, but the Total is set based on their higher overall season average.

In general, those 6.5 to 8 totals just happen to be the sweet spot for the UNDER since 2007. For example, just betting all games with a Total of 6.5 since 2007 has an ROI of 5.9% ... total=6.5 and H and season>=2007

A lot of the time it could simply be the starters era is inflated compared to their peripherals. Era generally not the best indicator of what we should expect from a starter moving forward. I do think ppl in general do tend to look at it though and most likely look to bet over in games w totals 8 or less when seeing starters with higher era,, I really think you are on to something here that will almost assuredly see positive gains all season.. great idea.

[TechnicalTrader]

The RoI on this:

H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2007

is 19.9% (dating back to 2013), after the all star break, compared to "only" 7.2% before the break:

AASB and H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2013

AASB=0 and H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2013

I am guessing that this has to do with playoff bound teams "tightening up" their D during the race, crappy teams tanking and the fact that prospects tend to get the most playing time after the break, when rosters expand. Just a theory though.

[barryt]

Just wondering if it works for( SDTDSERA plus o:SDTDSERA)>=total.
I.e. The sum being >= total rather than each starter >= total/2. Other restrictions still applying.
I don't know if you can query it that way or not. ( I need to learn this SDQL if I'm going to hang out here, downloaded the instruction book some time ago but never got around to it)

[barryt]

The RoI on this:

H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2007

is 19.9% (dating back to 2013), after the all star break, compared to "only" 7.2% before the break:

AASB and H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2013

AASB=0 and H and REG and 6.5<=total<=8 and total/2<=STDSERA<=5.25 and total/2<=o:STDSERA<=5.25 and 1.25<=STDSWHIP<=1.55 and 1.25<=o:STDSWHIP<=1.55 and season>=2013

I am guessing that this has to do with playoff bound teams "tightening up" their D during the race, crappy teams tanking and the fact that prospects tend to get the most playing time after the break, when rosters expand. Just a theory though.

After the ASB, the STDSERA are based on more starts than before the ASB and so the average data is more indicative.( accurate)??

[nfl_huskers]

I won't be at a computer until after games start tonight, so here are tonight's plays as they stand right now. Using 5Dimes odds and also showing SDQL odds. I usually like official plays closer to game time to make sure the system params are met.

Filtered System (13%ROI)

MIL @ STL: 7.5U -105 (5Dimes), -110 (SDQL)

COL @ SD: 7.5U -115 (5Dimes), -125 (SDQL)

CLE @ DET: 8.0U -115 (5Dimes), -110 (SDQL)


Unfiltered System (5% ROI)

Same as above, but add ...

TB @ MIA: 8.0U -105 (5Dimes), -110 (SDQL)


Good luck!

[nfl_huskers]

TB-Miami went under. Welcome huskers, very nice to see you here! There are some brilliant SDQL and MLB experts on these boards.

Hey Scrivero - good to see you on here too! Working on another NHL angle I'll need to show you for next season.

[nfl_huskers]

Good stuff. I watched your thread across the street but was a bit too busy here to follow or provide feedback. Good stuff and thanks for bringing it over here!

I am guessing the reason why the unders are generating a nice RoI is simply because the market is undervaluing these pitchers. I'll see what I can find.

Hey TT - I've followed many of your posts. Very impressive SDQL systems. I've been tracking your current thread with the line movements. Incredible returns. Looking forward to discussing ideas with you.

[nfl_huskers]

A lot of the time it could simply be the starters era is inflated compared to their peripherals. Era generally not the best indicator of what we should expect from a starter moving forward. I do think ppl in general do tend to look at it though and most likely look to bet over in games w totals 8 or less when seeing starters with higher era,, I really think you are on to something here that will almost assuredly see positive gains all season.. great idea.

Thanks for the positive feedback. I agree that the ERA is not the best indicator, and maybe bettors read too much into it. Let me know if you notice any other patterns in the results. I'm just getting started.

[TechnicalTrader]

Hey TT - I've followed many of your posts. Very impressive SDQL systems. I've been tracking your current thread with the line movements. Incredible returns. Looking forward to discussing ideas with you.

Like-wise. I've been watching your other thread for a few weeks and have been tailing myself. Very, very good work.

Throw some ideas at me in a PM, I'm looking forward to exchanging ideas and thoughts!

[nfl_huskers]

Just wondering if it works for( SDTDSERA plus o:SDTDSERA)>=total.
I.e. The sum being >= total rather than each starter >= total/2. Other restrictions still applying.
I don't know if you can query it that way or not. ( I need to learn this SDQL if I'm going to hang out here, downloaded the instruction book some time ago but never got around to it)

Hey barryt - we think alike. I had tried that same query back in early April, and it didn't impress me. Just ran it again with no filter on ERA and WHIP. The ROI drops from 5.8% to 1.7% across all SDQL seasons. Still positive for many plays, so there might still be an angle in there.

Possible that if one of the pitchers has a very low ERA and the other is very high, then more Under bets than Over are made antipating a 6-0 or 7-0 shutout type game, and the value is lost.

[TechnicalTrader]

Here's a little side project I've been working on (sorry for hijacking the thread) where I look at specific ace's performances and see if I can find a pattern. Kershaw is good, Kershaw is real, real good. I think we can all agree to that. What does Kershaw typically do after this "type of start", going up against a mediocre starter (which he will in SD)???

This is what he does:


6-0 SU
6-0, RL
3 games where he gave up 0 runs
in the other two, he gave up a combined 3 earned runs in 11 innings.

I will play the following on his next start:

LAD, SU
LAD, RL
SD TT, Under 2.5 (if I can get it)
Maybe even a player prop, Clayton under two earned runs (if I can get it)

[2daBank]

Thanks for the positive feedback. I agree that the ERA is not the best indicator, and maybe bettors read too much into it. Let me know if you notice any other patterns in the results. I'm just getting started.

Not sure how much help I'd be, you guys doing these sdql things are on another planet as opposed to how I go about things! Lol. I do find it incredibly interesting tho and think ya'll a asset to the forum ( I enjoy seeing the other ways ppl go about things!)..

One thing,, I believe I read you say you prefer to wait till closer to gametime to wager is that correct? I'm not 100% sure about this but I think you will find more often than not you will lose value by waiting with this type of play. Certainly the games where the pitchers era is higher than peripherals you can count on either paying more juice or losing a half run by waiting , you be far better off playing them overnight imho.

[TechnicalTrader]

Not sure how much help I'd be, you guys doing these sdql things are on another planet as opposed to how I go about things! Lol. I do find it incredibly interesting tho and think ya'll a asset to the forum ( I enjoy seeing the other ways ppl go about things!)..

One thing,, I believe I read you say you prefer to wait till closer to gametime to wager is that correct? I'm not 100% sure about this but I think you will find more often than not you will lose value by waiting with this type of play. Certainly the games where the pitchers era is higher than peripherals you can count on either paying more juice or losing a half run by waiting , you be far better off playing them overnight imho.

I am a huge student of the game. Played semi-pro ball for over 15 years and just recently retired from baseball. I've caught a few thousand innings and have served as an assistant coach several years. I'm not your typical numbers dork, I really, really love the game...but, but, but SDQL makes life sooooo much easier. Imagine how long it would take you to get the above stat on Kershaw. that took me about 20 seconds. Capping is so much easier with SDQL a "one" of your tools. It's never too late to learn it!

[2daBank]

I am a huge student of the game. Played semi-pro ball for over 15 years and just recently retired from baseball. I've caught a few thousand innings and have served as an assistant coach several years. I'm not your typical numbers dork, I really, really love the game...but, but, but SDQL makes life sooooo much easier. Imagine how long it would take you to get the above stat on Kershaw. that took me about 20 seconds. Capping is so much easier with SDQL a "one" of your tools. It's never too late to learn it!

I'm certainly curious. I'm actually a bit of a "number dork" myself, extreamly far from typical but none the less, the reason I love bases is cause I love capping it and it all numbers for me!! I'm basically computer illiterate tho and way I go about things far from normal for a numbers guy but I believe in the "science" of how I go about things and think my results have supported this over the years. I Really do appreciate you and the few others taking time to share your approaches as I said I love seeing how others go about this thing we do.

[doubledime]

I am a huge student of the game. Played semi-pro ball for over 15 years and just recently retired from baseball. I've caught a few thousand innings and have served as an assistant coach several years. I'm not your typical numbers dork, I really, really love the game...but, but, but SDQL makes life sooooo much easier. Imagine how long it would take you to get the above stat on Kershaw. that took me about 20 seconds. Capping is so much easier with SDQL a "one" of your tools. It's never too late to learn it!

TT, I didn't know they played baseball on your side of the pond. I thought only cricket.

Good luck today

[Scrivero]

Hey Scrivero - good to see you on here too! Working on another NHL angle I'll need to show you for next season.

Nice, I cant wait! And yea Im here nowadays because its just so much more stuff going on here. So many people offering a hand and helping with the systems.

[TechnicalTrader]

TT, I didn't know they played baseball on your side of the pond. I thought only cricket.

Good luck today

LMAO, my old buddy DD came in for the low blow!

[BarstoolProphet]

TT can you do a sdql for this:

When the game total is 7 or lower. How often is the final score determined by 1-2 runs ,either side.

Please provide link. I cant seem to get it to work. thx

[TechnicalTrader]

Here are your 1's and 2's

season = 2017 and total<=7 and margin=1,2,-1,-2

Just deduct them from total games played. Gotta run, later

[BarstoolProphet]

THx I think I might have found something that can avg 20+ units a year. Back testing now. But you would need to bet the dog to win by exactly one run. If that is avail at your book.

If it is avail, could someone please tell me what the odds are for LAA to win by exactly one run.

My book currently has it at +550, it is under run line alts like this.

LAA -1 +195
TIE +1 +550
SEA +1 -167

So If I understand this correctly if LAA wins by exactly one run the tie pays out.

I read it that way bc my book also has it in reverse. (Tie pays out if SEA wins by 1)

LAA +1 +115
TIE -1 +350
SEA -1 +130

[TechnicalTrader]

bet365 has it at +550 also:

https://www.bet365.com/?cb=103264292823433876#/AC/B16/C20525425/D19/E4080927/F19/P^48/Q^1/W^36

"winning margin"

[Scrivero]

I won't be at a computer until after games start tonight, so here are tonight's plays as they stand right now. Using 5Dimes odds and also showing SDQL odds. I usually like official plays closer to game time to make sure the system params are met.

Filtered System (13%ROI)

MIL @ STL: 7.5U -105 (5Dimes), -110 (SDQL)

COL @ SD: 7.5U -115 (5Dimes), -125 (SDQL)

CLE @ DET: 8.0U -115 (5Dimes), -110 (SDQL)


Unfiltered System (5% ROI)

Same as above, but add ...

TB @ MIA: 8.0U -105 (5Dimes), -110 (SDQL)


Good luck!

Aaaaand you hit everything again. What I do know is that I make my fade O/U plays first. Then I check if you have any plays that I dont already have. THen I play them. THen I check if sportsbetter21 has more games that I dont already have and I bet them. Went 5-0 tailing this way yesterday. One of my fades failed and the other one is still pending. Im probably just getting lucky with tailing the right people.

To make it short... please post these plays every day!

[Scrivero]

Oops, I thought the Colorado-game was over when I looked at it. But it wasnt, ended with 8 runs. Anyways, 3-1 for your plays.

[TechnicalTrader]

Great day. 3-1. SDQL strategies are killing it right now!

[TechnicalTrader]

LAA/SEA under 8.0
CWS/KC under 8.0

and

CWS/KC under 8.0

plays for today, correct?

[TempleoftheDog]

Hi TT,
I have a few questions on SDQL data MLB...
1) when a starter is 'red' = early exit during the match?
2) starter 'green' = ?
3) starter 'black' = ?
I believe on every SDQL we find the correct starter of both teams... the problem is I'm in europe and when teams change starter late in the day in usa, my 'totals' bet gone void.
Is there a stat of how many games have a starter change?
Is there a rule of how many hours before a match a team can change the starter?
I already apologize if my questions are stupid... only one year in the mlb world... and sorry for my english...
thank you

[TechnicalTrader]

I'm not 100% sure but I think green is a pitcher who qualified as having a quality start and red is bad start. Not 100% sure though

[TempleoftheDog]

thank you for the reply...
yes, I think you are right... if you put 'BS' appear a lot of red, 'QS' all green...
maybe there are something more...

have a nice day

[nfl_huskers]

LAA/SEA under 8.0
CWS/KC under 8.0

and

CWS/KC under 8.0

plays for today, correct?

Lots of questions at this point.

PHI/CHC under 7.0 is looking good for both systems. Just showed up on SDQL. One site has Eickhoff's ERA at 3.86, but most have it at 3.56. If total goes to 7.5, it's not a play at 3.56. Looks like it will hold 7.0 though.

LAA/SEA under 8.0 is a possible play on the unfiltered only. Total looks like it's heading for 8.5 though. Will watch.

CWS/KC under 8.0 is a possible play on the unfiltered only. SDQL has Quintana as the CWS starter, but I'm seeing Pelfrey everywhere else. Also, looks like SDQL thinks Karns' ERA is 4.83, but I'm seeing 6.26. These errors put the game on the filtered system, but it's not.

So ... in a holding pattern on official plays.