i think something isn't right with my hrr model guys. it's projecting too big of an edge and despite some success the results don't line up with the projected edge. rsigley taught me that when results don't align with expectations you need to take a step back and re-evaluate. i'm betting too many unders and under-projecting too many guys. that's a systematic bias and a sign something isn't right. i'm guessing it's most likely how i give credit to the runs+rbis to each player. probably not giving enough credit to the studs so lets take a step back and see if we can diagnose this.

It's hard though because if they offer you + odds against Harper you want to take it because all he does is hit bombs and doesn't get on base any other way. Not sure if it helps but I'm at ~43.5% unders and ~56.5% overs with HRR over 246 bets. And 54% unders in K's and 46% over in K's. Almost match up!

You could parse all of it if you want, but this data is pretty accurate

Over 0.5 Runs in First

Total: 6.5 (n=456) Over: 139 (41.89) Under: -139 (58.11)
Total: 7 (n=1341) Over: 110 (47.58) Under: -110 (52.42)
Total: 7.5 (n=2233) Over: 106 (48.45) Under: -106 (51.55)
Total: 8 (n=1852) Over: 101 (49.78) Under: -101 (50.22)
Total: 8.5 (n=2378) Over: -107 (51.77) Under: 107 (48.23)
Total: 9 (n=1803) Over: -118 (54.19) Under: 118 (45.81)
Total: 9.5 (n=979) Over: -111 (52.6) Under: 111 (47.4)
Total: 10 (n=404) Over: -138 (57.92) Under: 138 (42.08)
Total: 10.5 (n=298) Over: -124 (55.37) Under: 124 (44.63)

Over 1.5 Runs in First

Total: 6.5 (n=456) Over: 361 (21.71) Under: -361 (78.29)
Total: 7 (n=1341) Over: 298 (25.13) Under: -298 (74.87)
Total: 7.5 (n=2233) Over: 272 (26.87) Under: -272 (73.13)
Total: 8 (n=1852) Over: 262 (27.65) Under: -262 (72.35)
Total: 8.5 (n=2378) Over: 250 (28.55) Under: -250 (71.45)
Total: 9 (n=1803) Over: 221 (31.17) Under: -221 (68.83)
Total: 9.5 (n=979) Over: 241 (29.32) Under: -241 (70.68)
Total: 10 (n=404) Over: 168 (37.38) Under: -168 (62.62)
Total: 10.5 (n=298) Over: 173 (36.58) Under: -173 (63.42)

Odd or Even Final Score

Total: 6.5 (n=456) Even: 173 (36.62) Odd: -173 (63.38)
Total: 7 (n=1341) Even: 148 (40.27) Odd: -148 (59.73)
Total: 7.5 (n=2233) Even: 146 (40.57) Odd: -146 (59.43)
Total: 8 (n=1852) Even: 154 (39.31) Odd: -154 (60.69)
Total: 8.5 (n=2378) Even: 134 (42.77) Odd: -134 (57.23)
Total: 9 (n=1803) Even: 133 (42.87) Odd: -133 (57.13)
Total: 9.5 (n=979) Even: 149 (40.14) Odd: -149 (59.86)
Total: 10 (n=404) Even: 123 (44.8) Odd: -123 (55.2)
Total: 10.5 (n=298) Even: 118 (45.97) Odd: -118 (54.03)

Early TB at The Greek (Olympic sportsbook for the oldtimers here)

Alonso -115 over Hoskins
Abreu -115 over H. Dozier
Rosario -130 over Galvis

Total Bases:
Y. Alonso +103 over Soler (BM)
E. Rosario -130 over F. Galvis
G. Springer -110 over K. Davis
M. Betts -130 over B. Gardner
D. Peralta +105 over R. Acuna Jr.
P. Goldschmidt +115 over C. Yelich
B. "Gonna hit a dinger today" Harper -115 over M. Conforto

Well my bad Harper isn't the best example because he is more of a true hitter a better example is Joey Gallo who we fade all the time. This works well especially for lefties who are struggling against the shift

K's:
Jake Arrieta o5 -114
Corbin Burners u5.5 +110
Michael Wacha Flocka Flame o5 +120
Walker Buehler o5.5 -114

Oh no 50% of Harpers TB come from HR's, I don't think he's that good actually. Also he walks a lot. But I think he has a really god matchup today and I know Arrieta has given up a lot of homeruns, with the weather and the park I think he can keep them contained #narrativestreet

He basically either hits a double or a home run.

Something we should consider as a team. One key component of HRR is the second R, RBI's. Are elite players penalized in that they tend to be intentionally walked or pitched around when other people are on bases?

Early HRR
Realmuto u2.0 -114
Aguilar u2.0 -114
DeJong u2.0 -114
A. Gordon u2.0 -114
Moncada u2.0 -130
Pollock u2.0 -114

It's a good point but also the second R is the only stat where a hitter can accumulate more than one of that stat in a single AB so it balances out

Tim Anderson u2 -130
Goldschmidt o2 -114
Abreu u2 -114
J. Aguilar u2 -114
Eloy Jimenez u -114 (900 star play)
Cody Bellinger o2 -114
Harper o2 -114 (dinger theory)

I'm noticing a disturbing trend with BetBuilder and K's. They're not posting every pitcher.

And almost every HRR H2H posted has one person not playing. Nice.

Lucas giolito o5 K +100 (Jazz)

We need these tracked in the thread!!

yes the position for run scored in first analyst is open to anyone please join us

Too hard because bet sizing is so important and it's usually better to bet close to game time, but one that looks good at least from the chart is

SLN/MIL no score 1I +120

fair is +111 so that should be a 1.18% of bankroll bet. This is where you run a risk tho. It's -113 on the over, is it going to 10? Unlikely, but if it's -125 does that effect it? That's where straight derivatives get tricky. Long run it shouldn't matter because you'll get as many -125 on the over as -125 (which you'll probably overbet) on the under (which you'll underbet probably), but if I spent more than 17 minutes doing it I might subset it into probabilities and create cut off, etc. but that sounds like too much work for little gain. Like I'm looking at TOR/MIN HRE u27 -105 at Westgate right now and wanting to bet it, but I'm thinking a fair total is more 9 so I'll lay off and see what happens to the total throughout the day. Maybe I'll miss out, maybe I'll stop myself from making a -EV bet.

Lets try this one again Votto HRR +110 over J. Turner (maybe he won't chicken out this time and not start)

TB: Alex Gordon -103 over Eloy Jimenez (Bookmaker)

False alarm with the BetBuilder K's, thank you Ari for posting more
Sonny Gray u5.5 -114
Carrasco u6.5 +110
Trevor williams u5 -114
Hellickson o3.5 -114
Hamels u6 +110
Odorizzi u5.5 +110
Lance Lynn o4.5 +110
Montas o3.5 -114
Miley o4 +127

I really need to get in on this bet builder business

some great discussion today.

rsigley do you use your projected lines when estimating the fair value of props like run in first and rhe? or do you use the current betting line? i can see an argument for either approach.

we've also opened up a more junior level role, run in first apprentice so those who don't feel 100% comfortable diving into a more senior level analyst role can still contribute.

Great question Tom I had the same. I’ll take the junior role here and guess he looks at both and tries to judge. Probably uses his projections as a lean/tiebreaker


Sent from my iPhone using Tapatalk

I have my own predictions which are based on projected lineups, weather, etc. usually they align pretty well with the market - only like 20% of my bets are totals vs 80% ML on full game stuff.

By the time these type of props come out all the big movements in the lines happen, any other movements after are usually due to something like someone not being in a lineup or weather changes, lineup changes, etc.

So in that example I have it closer to 9 (8.9 is my prediction), so I'll treat it at 9, but if all of a sudden steam comes in on under or there's a lineup change I might change.

Also congrats on our first RSIF prop bet, easy +1.2 and then cardinals score a bunch in the 2nd

harper only has two more chances to get a dinger

I have a no0b, bottom feeder question. If someone can answer it for me Ill happily provide my NHL player bets for tonight if thats something that tickles your fancy...

Which of these bets is better/more correlated?

PARLAY (2 TEAMS)
[927] TOTAL o8½EV (PIT PIRATES vrs DET TIGERS)
( T WILLIAMS -R / S TURNBULL -R )
[927] PIT PIRATES -1½+150
( T WILLIAMS -R / S TURNBULL -R )

or

PARLAY (2 TEAMS)
[1927] TOTAL o4½-105 (1H PIT PIRATES vrs 1H DET TIGERS)
( T WILLIAMS -R / S TURNBULL -R )
[1927] 1H PIT PIRATES -½+125
( T WILLIAMS -R / S TURNBULL -R )



In other words, are 1h baseball CP's better than full game? Intuitively it seems to me that 1h would be better. But the ratio of run line to total is closer in full game than in it is for 1h. Wondering if anyone can conclusively answer this for me, and again, I'll provide some +EV bets in return if you guys want.

Late game HRR

Devers u2.0 -130
Haniger u2.0 -114
D. Santana u2.0 -130
Goodrum u2.0 -130
Zimmerman u2.0 -114
Escobar u2.0 -114
T. Hernandez u2.0 -114
Smoak u2.0 -114

I glanced up at the television in the Wynn Sportsbook and saw a homerun and was like Harper got it.

but it was Scott Kingery

lots of ways to do this. you gotta have a database though. no way around it. risgley you are right can't hammer this one home enough. retrosheet will work for this problem since we aren't calculating if it's +ev vs betting lines we just want to compare correlation. categorize each game result into each of the scenarios you are interested in (-1.5 rl + over, -1.5 rl + under, +1.5 rl + over, +1.5 rl +under) and see what percentage of the whole they fall into each category. do the same for the 5i. then just tally 'em up and see how often each group occurs compared to the other. great question and opportunity to get started building out a database and working through a problem.

Lucas Giolitto gets injured after 2.2 and 5K's

He's currently 1-2 on the batter.

If the relief pitcher gets the third strike does Giolitto get the 6th K and we win? Or we pushing

Problem with 5i is the range of the totals is so small. So do you separate it based on implied probability? So break 3.5 into 3.25, 3.5, etc. like they do in soccer?

Edit: Burr struck him out and they gave the K to him and not Giolito. What a scam way to push a bet we crushed.