How can we use this trend?

[letsgohoosiers]

Bet on any team and the UNDER with a starting pitcher who's ERA at the end of the game is below 4.2

http://sportsdatabase.com/mlb.py/query?text=SERA<4.2 and month&sort=query_header

<table style="width: 739px; height: 91px;" id="sortable_table" bgcolor="#ffffff" border="0" cellpadding="2"><thead><tr><th>Record
W-L (marg, % win)</th> <th>
$ On</th> <th>Over/Under
O-U-P (marg, % over)</th> <th>
$ Under </th> <th name="query_header">season</th> </tr></thead><tbody> <tr bgcolor="#ffffff"> <td align="center">303-187 (1.27, 61.8%)</td> <td align="center"> 10965 </td> <td align="center">196-272-21 (-0.64, 41.9%)</td> <td align="center"> 5515 </td> <td align="center">2009</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">1502-1068 (0.90, 58.4%)</td> <td align="center"> 31245 </td> <td align="center">1051-1401-116 (-0.20, 42.9%)</td> <td align="center"> 23210 </td> <td align="center">2008</td> </tr> <tr bgcolor="#ffffff"> <td align="center">1655-1216 (0.75, 57.6%)</td> <td align="center"> 32135 </td> <td align="center">1210-1505-152 (-0.01, 44.6%)</td> <td align="center"> 16150 </td> <td align="center">2007</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">1589-1324 (0.45, 54.5%)</td> <td align="center"> 14975 </td> <td align="center">1287-1482-141 (-0.00, 46.5%)</td> <td align="center"> 5855 </td> <td align="center">2006</td> </tr> <tr bgcolor="#ffffff"> <td align="center">1445-1067 (0.78, 57.5%)</td> <td align="center"> 23995 </td> <td align="center">982-1370-160 (-0.30, 41.8%)</td> <td align="center"> 28395 </td> <td align="center">2005</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">1308-931 (0.93, 58.4%)</td> <td align="center"> 23640 </td> <td align="center">937-1210-92 (-0.15, 43.6%)</td> <td align="center"> 17930 </td> <td align="center">2004</td></tr></tbody></table>
I found this trend using SERA<4.2 and season

However, it's fairly "duh" because if they pitch well you will have a good chance at winning, predicting who pitches well is the tough part.

Also, the majority of the money is made in April, because that's when the ERA is highly unstable.

http://sportsdatabase.com/mlb.py/query?text=SERA<4.2 and month&sort=query_header

<table style="width: 753px; height: 77px;" id="sortable_table" bgcolor="#ffffff" border="0" cellpadding="2"><thead><tr><th>Record
W-L (marg, % win)</th> <th>
$ On</th> <th>Over/Under
O-U-P (marg, % over)</th> <th>
$ Under </th> <th name="query_header">month</th> </tr></thead><tbody> <tr bgcolor="#ffffff"> <td align="center">169-110 (0.89, 60.6%)</td> <td align="center"> 5605 </td> <td align="center">112-163-4 (-0.63, 40.7%)</td> <td align="center"> 3900 </td> <td align="center">10</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">1219-1000 (0.55, 54.9%)</td> <td align="center"> 6260 </td> <td align="center">969-1136-111 (0.08, 46.0%)</td> <td align="center"> 5475 </td> <td align="center">9</td> </tr> <tr bgcolor="#ffffff"> <td align="center">1289-1022 (0.63, 55.8%)</td> <td align="center"> 13480 </td> <td align="center">1029-1176-106 (0.14, 46.7%)</td> <td align="center"> 3480 </td> <td align="center">8</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">1146-929 (0.65, 55.2%)</td> <td align="center"> 9460 </td> <td align="center">943-1038-93 (0.24, 47.6%)</td> <td align="center"> -95 </td> <td align="center">7</td> </tr> <tr bgcolor="#ffffff"> <td align="center">1186-887 (0.75, 57.2%)</td> <td align="center"> 19055 </td> <td align="center">846-1116-108 (-0.13, 43.1%)</td> <td align="center"> 17970 </td> <td align="center">6</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">1403-1014 (0.74, 58.0%)</td> <td align="center"> 29430 </td> <td align="center">963-1303-151 (-0.30, 42.5%)</td> <td align="center"> 23860 </td> <td align="center">5</td> </tr> <tr bgcolor="#ffffff"> <td align="center">1378-826 (1.27, 62.5%)</td> <td align="center"> 53000 </td> <td align="center">797-1295-109 (-0.81, 38.1%)</td> <td align="center"> 41610 </td> <td align="center">4</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">12-5 (2.00, 70.6%)</td> <td align="center"> 665 </td> <td align="center">4-13-0 (-0.85, 23.5%)</td> <td align="center"> 855 </td> <td align="center">3</td></tr></tbody></table>
Fading a pitcher's ERA after the game above 4.8 works as well, in addition to playing the OVER.

http://sportsdatabase.com/mlb.py/query?text=SERA>4.8 and season&sort=query_header

<table style="width: 752px; height: 120px;" id="sortable_table" bgcolor="#ffffff" border="0" cellpadding="2"><thead><tr><th>Record
W-L (marg, % win)</th> <th>
$ Against</th> <th>Over/Under
O-U-P (marg, % over)</th> <th>
$ Over </th> <th name="query_header">season</th> </tr></thead><tbody> <tr bgcolor="#ffffff"> <td align="center">195-312 (-1.31, 38.5%)</td> <td align="center"> 11220 </td> <td align="center">322-161-19 (2.13, 66.7%)</td> <td align="center"> 15265 </td> <td align="center">2009</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">525-950 (-1.53, 35.6%)</td> <td align="center"> 34870 </td> <td align="center">816-584-73 (1.31, 58.3%)</td> <td align="center"> 17945 </td> <td align="center">2008</td> </tr> <tr bgcolor="#ffffff"> <td align="center">513-890 (-1.34, 36.6%)</td> <td align="center"> 29805 </td> <td align="center">758-575-63 (1.27, 56.9%)</td> <td align="center"> 12445 </td> <td align="center">2007</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">537-833 (-1.09, 39.2%)</td> <td align="center"> 22735 </td> <td align="center">791-519-57 (1.65, 60.4%)</td> <td align="center"> 22590 </td> <td align="center">2006</td> </tr> <tr bgcolor="#ffffff"> <td align="center">552-901 (-1.34, 38.0%)</td> <td align="center"> 22595 </td> <td align="center">801-582-70 (1.20, 57.9%)</td> <td align="center"> 16610 </td> <td align="center">2005</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">693-1097 (-1.18, 38.7%)</td> <td align="center"> 29425 </td> <td align="center">994-716-80 (1.40, 58.1%)</td> <td align="center"> 20640 </td> <td align="center">2004</td></tr></tbody></table>
However, again is this useful because it's largely based on numbers in early months when we don't know how good/bad pitchers are?

http://sportsdatabase.com/mlb.py/query?text=SERA>4.8 and month&sort=query_header

<table style="width: 749px; height: 120px;" id="sortable_table" bgcolor="#ffffff" border="0" cellpadding="2"><thead><tr><th>Record
W-L (marg, % win)</th> <th>
$ Against</th> <th>Over/Under
O-U-P (marg, % over)</th> <th>
$ Over </th> <th name="query_header">month</th> </tr></thead><tbody> <tr bgcolor="#ffffff"> <td align="center">53-102 (-1.39, 34.2%)</td> <td align="center"> 4735 </td> <td align="center">85-65-5 (1.25, 56.7%)</td> <td align="center"> 1390 </td> <td align="center">10</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">456-677 (-1.05, 40.2%)</td> <td align="center"> 8720 </td> <td align="center">594-487-46 (1.11, 54.9%)</td> <td align="center"> 6605 </td> <td align="center">9</td> </tr> <tr bgcolor="#ffffff"> <td align="center">415-678 (-1.31, 38.0%)</td> <td align="center"> 17115 </td> <td align="center">567-478-48 (1.10, 54.3%)</td> <td align="center"> 4065 </td> <td align="center">8</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">446-642 (-0.98, 41.0%)</td> <td align="center"> 10645 </td> <td align="center">608-431-48 (1.42, 58.5%)</td> <td align="center"> 13650 </td> <td align="center">7</td> </tr> <tr bgcolor="#ffffff"> <td align="center">452-753 (-1.34, 37.5%)</td> <td align="center"> 22520 </td> <td align="center">687-470-47 (1.40, 59.4%)</td> <td align="center"> 17570 </td> <td align="center">6</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">611-1010 (-1.27, 37.7%)</td> <td align="center"> 33680 </td> <td align="center">874-655-85 (1.24, 57.2%)</td> <td align="center"> 15210 </td> <td align="center">5</td> </tr> <tr bgcolor="#ffffff"> <td align="center">579-1111 (-1.63, 34.3%)</td> <td align="center"> 52420 </td> <td align="center">1058-547-83 (1.99, 65.9%)</td> <td align="center"> 46535 </td> <td align="center">4</td> </tr> <tr bgcolor="#e6e6e6"> <td align="center">3-10 (-2.92, 23.1%)</td> <td align="center"> 815 </td> <td align="center">9-4-0 (3.42, 69.2%)</td> <td align="center"> 470 </td> <td align="center">3</td></tr></tbody></table>
It would be MUCH easier to figured this out and shoot it down if they simply pulled the ERA of the starter BEFORE the game in their query as opposed to AFTER the game.

 

[TheJackal]

SERA is the starter's ERA for that particular game (meaning it hasn't happened yet when you place your wager). You need to find a trend based on previous games which would include a query with p:SERA (buut again that will just be the ERA for a single game.) You will need to use summative methods to get the ERA for a series, season, etc if yo uwant to build a trend of more than a single performance by a starter.

You also probably have a lot of double counted games where both starters had an ERA under 4.2 for that particular game. (remember there are two records in the database for each game, one for each team so when dealing with O/U you have to be careful not to double count them)

Good start though. I'll keep an eye on where this is going.

 

[YourAllAmerican]

predicting who pitches well is the tough part.

That pretty much sums up the whole challenge of baseball right there. Just about any system would work well if you could predict ahead of time who would pitch well.

 

[letsgohoosiers]

SERA is the starter's ERA for that particular game (meaning it hasn't happened yet when you place your wager). You need to find a trend based on previous games which would include a query with p:SERA (buut again that will just be the ERA for a single game.) You will need to use summative methods to get the ERA for a series, season, etc if yo uwant to build a trend of more than a single performance by a starter.

You also probably have a lot of double counted games where both starters had an ERA under 4.2 for that particular game. (remember there are two records in the database for each game, one for each team so when dealing with O/U you have to be careful not to double count them)

Good start though. I'll keep an eye on where this is going.

There are TONS of double counted games, and just no way to get rid of them which not necessarily ruins the analysis, but skews it. I think if I was betting I would lay off ML plays with multiple qualifiers but play totals.

I'm not sure SERA is for that particular game, rather his season ERA after the game. One of the SERA>4.8 (fade) plays yesterday was Niemann for TB. He pitched 8 innings giving up 3 ERs, so that's not a >4.8 ERA for the game. I believe they're pulling the ERA query from Covers' ERA after the game which is his season ERA including the game he just pitched.

I just can't think Baseball is as simple as playing good pitchers and the under as well as fading bad pitchers and playing the over, but it looks that way.

I'll start tracking plays to see how this goes. Here's today's ERA games:

ATL J Jurrgens 2.06 ERA -180
COL-ATL UNDER 9.0 -120
CIN J Cueto 1.93 ERA +134
PHI-CIN UNDER 8.0 -103
FLA J Johnson 2.50 ERA -160
ARI-FLA UNDER 8.5 -120
CWS M Buehrle 3.00 ERA -115
MIN-CWS UNDER 8.5 -104
LOS C Billingsley 2.30 ERA -135
NYM-LOS UNDER 8.0 -102

Odd that Cueto is drawing plus money at home, especially against a pitcher with a losing record and a 5.0+ ERA in Hamels. I'll look at high ERAs later, need to make a grocery run.

 

[letsgohoosiers]

Hmmm

I just clicked on the link, and the numbers are completely different.

I wonder what happened.

 

[letsgohoosiers]

Now they're pulling the ERA for the game, must have been an error. Glad I got the results before the changed it.

 

[letsgohoosiers]

ATL J Jurrgens 2.06 ERA -180----WINNER
COL-ATL UNDER 9.0 -120--------PUSH
CIN J Cueto 1.93 ERA +134-------LOSER
PHI-CIN UNDER 8.0 -103---------WINNER
FLA J Johnson 2.50 ERA -160-----LOSER
ARI-FLA UNDER 8.5 -120---------WINNER
CWS M Buehrle 3.00 ERA -115----WINNER
MIN-CWS UNDER 8.5 -104--------WINNER
LOS C Billingsley 2.30 ERA -135---WINNER
NYM-LOS UNDER 8.0 -102--------PUSH

6-2-2 (+2.4 units) today playing low starter ERAs and the unders on the POSTED plays, the overall system did much better.

I went through the results today and tried to note the difference between starting ERA and ending ERA. On average the difference was .23, with all but Dontrelle Willis having their ERA change less than a point.

(Note: If anyone knows how to post tables from Excel into posts that keep the formatting let me know - until then I'll just take screenshots and past the image)

 

[TheJackal]

hoosiers-

There has been a ton of confusion over SERA on a very regular basis. SERA is indeed the ERA for the starter in that particular game. (It is just a shortcut for 'starter earned runs*9/starter innings pitched')

Joe has recently decided to add two new shortcuts in order to try and clear up the confusion around starter era. You can now use these two parameters:

YTDSWHIP is the year to date starter's WHIP
YTDSERA is the year to date starter's ERA

 

[letsgohoosiers]

hoosiers-

There has been a ton of confusion over SERA on a very regular basis. SERA is indeed the ERA for the starter in that particular game. (It is just a shortcut for 'starter earned runs*9/starter innings pitched')

Joe has recently decided to add two new shortcuts in order to try and clear up the confusion around starter era. You can now use these two parameters:

YTDSWHIP is the year to date starter's WHIP
YTDSERA is the year to date starter's ERA

Well I understand that's what it measures. And if you click on the link now, it indeed measures that. And the numbers are significantly different (better) than the numbers posted in this thread.

When I was experimenting with the database 2 days ago, it wasn't measuring the starter's ERA for the game, rather it was pulling the cumulative season ERA at the END of the game for the starter. THAT info is useful IMO, because the ERA of the starter doesn't change a significant amount once the season gets rolling, and you can profit big time off of it.

I went back and looked through games that the initial query pulled to make sure it was right. As I said, Niemann from TB was showing up as an example in the SERA>4.8 query before they fixed it (must be reading this post). He pitched 8 innings giving up 3 ERs, so that's not a >4.8 ERA for the game. They were pulling the ERA query from Covers' ERA after the game.

I have a full card today, I'll track this for a week and if it's successful I'll post plays.

 

[letsgohoosiers]

5/20 results

 

[letsgohoosiers]

5/21 data

 

[Funnyb25]

what the hell is going on here?

 

[letsgohoosiers]

what the hell is going on here?

I ran a query on SportsDatabase.com pulling SERA<4.2 (starter's ERA less than 4.2 after the matchup), and the results are the first and second table tables in the first post. I also ran a query pulling SERA>4.8 (starter's ERA more than 4.8 after the matchup) and the results are in the 3rd and 4th tables.

The results were very good, in 2004 it was up 916.35 units, 2005 it was up 915.95 units, in 2006 it was up 661.55 units, in 2007 it was up 905.35 units, and in 2008 it was up 1072.7 units. In 2009 it is up 429.65 units so far, which leads me to believe there's still ~500 units in profit to be made.

When you pull the query now, it shows the ERA of the starter in the game as opposed to the ERA of the starter after the game, which is why if you click on the link at the top it leads you to a query with different and much more profitable numbers. I am tracking opening and closing ERA each day to see if how much of a cushion I should give myself when selecting plays.

I'll post picks soon for you degenerates, I promise, I just need to formulate the gameplan. I'm getting a bit tired of scalping.

 

[Funnyb25]

will you post them on this thread or u gunna create a new one??

 

[Funnyb25]

what if 1 pitchers era is say 2.08 and the others is a 2.94.... do u take the 2.08 and the under or just the under? likewise a 5.02 era vs a 5.60 era.. do u take the 5.02 era or just the over??

 

[letsgohoosiers]

what if 1 pitchers era is say 2.08 and the others is a 2.94.... do u take the 2.08 and the under or just the under? likewise a 5.02 era vs a 5.60 era.. do u take the 5.02 era or just the over??

I don't play multiple qualifiers. If both ERAs are over 4.8, I don't play either team, I put 2 units on the OVER. If both ERAs are under 4.2, I don't play either team I put 2 units on the UNDER.

If there is a 6.3 ERA for one team and a 3.1 ERA for the other, I place 2 units on the team with the 3.1 ERA starter and don't play the over/under.

If there is a pitcher with a 2.7 ERA and another pitcher with a 4.6 ERA, I would play the 2.7 ERA and the UNDER for 1 unit each (2 units total).

The only time I don't have 2 units on the game is if both pitchers have an ERA between 4.2 and 4.8, in that case I don't bet on the game.

Again, I want to test out what number I should be playing and what number I should be fading before I go ahead and post picks. There are some obvious games and I can post those. Others the ERA is something like 5.1 or 4.0 and I'm not sure if I want to play those games or not. In addition a lot of these starters only have 1-3 starts at this point and it wouldn't be wise to play them as their ERA is still highly volatile and remember we're trying to play the ERA after the game.

Here's two for tomorrow:

TOR - Halladay -155 (2 units)
PIT - Duke +115 (2 units)

 

[letsgohoosiers]

OK I'll give you more

NYM - Santana -118 (2 units)
CIN-CLE OVER 10.5 (+107) (2 units)
TAM-FLA OVER 10 (-107) (alternatively you can take OVER 10.5 for +124) (2 units)

 

[TheJackal]

Basically here are the queries for the theory you are talking about now that they have corrected the SERA issue and added the new YTDSERA parameter. I don't see the edge you are talking about.

Here you say take the over [both starters have >4.8 ERA prior to the game]($ OVER column):
http://sportsdatabase.com/mlb.py/qu...o:YTDSERA>4.8 and site=home&sort=query_header

Here you say take the under [both starters have <4.2 ERA prior to the game] ($ UNDER column):
http://sportsdatabase.com/mlb.py/qu...o:YTDSERA<4.2 and site=home&sort=query_header


Can you clarify the team selection again?
Is it one team needs <4.2 and it also needs to be < than the opponent starter
OR
is it one team <4.2 and the other over 4.8?

 

[letsgohoosiers]

Basically here are the queries for the theory you are talking about now that they have corrected the SERA issue and added the new YTDSERA parameter. I don't see the edge you are talking about.

Here you say take the over [both starters have >4.8 ERA prior to the game]($ OVER column):
http://sportsdatabase.com/mlb.py/qu...o:YTDSERA>4.8 and site=home&sort=query_header

Here you say take the under [both starters have <4.2 ERA prior to the game] ($ UNDER column):
http://sportsdatabase.com/mlb.py/qu...o:YTDSERA<4.2 and site=home&sort=query_header


Can you clarify the team selection again?
Is it one team needs <4.2 and it also needs to be < than the opponent starter
OR
is it one team <4.2 and the other over 4.8?

Remember, the queries I pulled initially (that they later fixed) were queries of the SERA AFTER the games, not the YTD SERA BEFORE the games (numbers are lower as you would expect), and not the SERA for the game itself (numbers were ridiculously higher).

I figured this out by going back into the history of the data and clicking through games and looking at numerous box scores. There were pitchers who had 5.0+ ERAs before the game that appeared on the less than 4.2 ERA list. There were pitchers that had 5.0+ ERAs for the game that appeared on the less than 4.2 list.

It makes sense that my numbers are inflated for April in this query, as you'd practically be perfect in week 1 outside of games where the results are changed due to the bullpen.

Team selection is something I'm playing with for a week or so before I am positive of what numbers I am doing. I know that I want the ERA after the game to be below a 4.2, or above a 4.8, to play/fade.

I want to FADE the pitcher and play the OVER on a starter who will have an ERA above 4.8 after the game. This is easily calculated, as you can add 9 innings of scoreless ball and recalculate ERA - if it's still above 4.8 it's a play.

I want to PLAY the pitcher and play the UNDER on a starter who will have an ERA below 4.2 after the game. This is a bit more difficult to calculate because you don't know how many innings they'll last, or how many runs they'll give up. A good measure would be to add 7 runs and recalculate ERA - if it's below 4.2 it's a play.

An ideal situation would be to have a starter under 4.2 and the opposing starter over 4.8. You simply put 2 units on the starter under 4.2.

If you both starters over 4.8, you play the OVER for 2 units. Likewise if they're both under 4.2, play the UNDER.

Then there's the situation when they're between the 4.2 and 4.8 range. I'm not certain what I want to do with that yet. I could simply use the data of the pitcher who isn't in that range and play/fade him and play the O/U, or just not play these games. I'll see how profitable this is while I'm still in beta

 

[letsgohoosiers]

5/22 results

I shouldn't have over thought this, starting pitchers with an ERA before the game under 4.2 went 6-2 and were 8-0 on the under.

Starting pitchers with an ERA before the game over 4.8 were 9-7 on the fade, but 5-12 on the over.

Would have netted me a 28-21 record if I kept it simple, but I don't do it that way. Also that's proven not to be successful in the long term.