Reverse Middle Scalps: Putting the House Percentage in Your Favor.

[mikren]

That's pretty close. Actually for a 20 cent middle (splitting 10 cents) as I used in my example your break even would be 1 in every 11 games (or 9.1%) because you win 10 cents 10 times and lose 100 once.

For a 25 cent middle as you used in your example your break even would be 1 in every 9 games because you win 12.5 cents 8 times and lose 100 once. 1 in 9 represents a breakeven of 11.1%. If you are assuming 10% pushes, that's not much of an edge, and a 20 cent middle is a long term loser at that assumed percentage. At 7.5% its a pretty good edge.
I think that whoever mentioned you need to look at the percentage of games that actually push at a given number as opposed to looking at the percentage of all games that fall on a given number is correct, and I would want to know those numbers before I committed myself to such a strategy. It would seem like anything 30 cents and above would be a solid edge though. Thanks for sharing the info.

[reno]

Mikren, the reverse middle/scalp would be between 9.5 and 10, not 9.0 and 10.

[sportshobby]

Boomer, this may help with your money management question:

Most players will tell you to increase your bet size when winning, and reduce it when losing. Yet Col (Ret) J.R. Miller demonstrates in his good book on NFL handicapping, that this strategy actually INCREASES the win percentage that one must maintain to break even. In other words, it makes it tougher.

The opposite strategy is doomed for failure (press when losing). So it seems that a more even bet size strategy is in order.

The Kelly criteria made a lot of hoopla a few years back. It derived the mathmatically optimal bet size, expressed in percent of bankroll, to maximize expected winnings. Problem is, it assumes that one bets one wger at a time, sees the result, then determines the next bet size. And it assumes that the same odds are played regardless of the bet size (usually not true).

I won't submit the mathmatical evidence, but I will give you the result of my study of this. Basically one should risk 2-5% of your bankroll on each wager, depending on the win percentage (from 55-60& for 11:10 wagers). It's pretty rare that a player really has a handle on what his wxpected win percentage is. But it is important to try, and to maintain records to periodically update that estimate. One should also avoid the upper end of the range (5%) until there is really some evidence that supports a high win percentage (60% ATS is almost a god).

[Boomer]

Sportshobby,

Doesn't the fact that I'm wagering sides of the bet make the wager size a whole lot more tricky. I mean if I bet three dimes on over 10, under 9.5 it definitely isn't the same as a gambler who wagers 3000 on the over, right.

That's why I'm having a difficult time knowing what is the right size to wager on these things. Any help from anyone would be appreciated. Again let's talk in terms of a hypothetical $10,000 bankroll. Wagering $200 on each side would give me a profit of $20-$30 and I know I want to bet more than that.

[sportshobby]

Boomer, you're right. It's a bit different for those reverse middle scalpe. If the probability of getting middled (losing big) is very small, your wager size could probably be 20% of bankroll. I've never analyzed this kind of situation.

[pay2play]

Sportshobby
I very much agree with your comments on Kelly. As a practical matter it has little value, I think. And in the heat of scalping/middling mania 10 minutes before post, fuggedaboutit. But the idea that ones bets should vary by win rate AND expectation is valueable, anyway.

BOOMER
Unlike the typical sportsbet, you're going to have a very high win pct, in the range of 90 percent. But the losses, of course, can be catastrophic if you hit any kind of streak. And as an experienced gambler I'm sure you know you'll always hit a losing streak that seems to defy the odds ESPECIALLY if you aren't prepared for it. I seem to recall a thread about some poor guy losing 11 straight accidental positions once(just teasing, Reno).
I mean, if you had a system dependent upon a roulette wheel not coming green twice in a row, you'd have no shot, right? We all know how it works.
If you're going to be losing 50% of your bet 10% of the time...damn, I forget how to make that calc.
All you're really worried about is disaster. You want protection against a bad run. You're bridgejumping, to borrow a racetrack phrase.
I'd say the best approach is to give it a separate small BR (because this is an experiment on your part) but make surprisingly large (relative to the seperate BR) bets. Just remember that the odds are you'll lose three in a row sometime this summer and it's going to...ah, you know how it'll feel.

[Cons]

Reno, isn't the hierachy the other way around?

I mean if you're a winner, you get worse lines, you constantly get put on hold, it takes longer for your bets to get accepted, often the line quoted will somehow change (never in your favour), and in the end you get thrown out and told to stay out.

If you are a loser, you are treated like a king. And the more you lose, the more valuable you are to those sportsbooks.

[Ronbets]

The 3 tier Hindu caste system reno is referring to has evolved from 2. Previously, anyone that paid markers(BMs or players) was considered a First Class passenger and everyone else was stowed with the baggage.
Today scalpers are even scoffed at on Wall Street. Computer driven buy/sell programs(arbitrage) have been asked to suspend activities on occasions(selloffs&triple witching days). Your right about that Rodney D. quote.

Sportshobby,
The late Huey Mahl was a strong proponent of the Kelly Criterium System. He didnt starve.

pay2pay,
The term you used EV (Expected Value) gives you away as a sharp pokerplayer. Robert Walker (Segundos pal?) of the Mirage fame did a study on the % of totals landing (all sports). One of his cohorts told me that his hold was higher moving the money rather than moving the #.
I agree with most of your math. However, bridgejumping might be too harsh a phrase for Boomers activity. Laying 30-1 with Mike Tyson over Buster Douglas is more fitting to the term. Keep up the great posts.--Ron

[Boomer]

A while there was a post about Murphy's Law, going to sleep and a team coming back and a whole bunch of other gambling laws.

Let's add "when trying out a new strategy, the worst possible scenario will occur early on to let you know you have no clue what you're doing."

As Tampa Bay/Philadelphia lands exactly on 9 and costs me 3 dimes.

[freak]

Based on your initial example......

If the game landed on 10 you lost $644....

Thus meaning a player betting for that $500 unit to make $50-$60 per bet, would have to win at least 13 $50 reverse scalps to break even for that 1 loss or 11 $60 to make $6.....

Now unless you have major money you can try to recoup it a lot quicker, but then again why risk all the money for $50- $60.....

Knowing you boomer, you are not a $500 reverse scalper, so your risk is a ton more....

Is it really worth the risk ???

To me I say no.....

Then again, I don't bet the amounts you do, or are capable of.

Continued success my friend.

Not beating ya down, just doing the math on recovering the loss is astonishing.

[Segundo]

Phila? What about the NYY and CIN games? This kind of stuff always happens. The only explanation of these totals landing like this today is that at least a handfull of people took your advice and fired away and now they DEFINITELY think you are the Devil incarnate; they can't get any good advice from you Boomer, what's up? (This was supposed to be funny, I'm just not that good at making jokes).

Ronbets, could you find out how Walker did his analysis? Ie. the specific math, theory involved? Thanks.

[reno]

Ronbets, I heard that Huey died poor. And a major reason was that the only thing he loved more than discussing Kelly money- management was the bottle. I'm so choked up thinking about it, I think I'll go pour myself another shot of Johnny Walker Black Label...

[AussieVamp2]

ok, so take your Kelly Criterion Deity handicapper - the 60% bloke

he finds 15 games that fit his strategies, whatever they may be, or 20, or 25 on a day

what does he do?

[AussieVamp2]

Ok, a look at the conditional totals thing from last year - this is put together fast from covers.com, but may be a reasonable guide :-

T O U P All O% U% P%
6 1 0 0 1 100.0% 0.0% 0.0%
6.5 2 1 0 3 66.7% 33.3% 0.0%
7 14 4 1 19 73.7% 21.1% 5.3%
7.5 30 29 0 59 50.8% 49.2% 0.0%
8 61 45 8 114 53.5% 39.5% 7.0%
8.5 161 135 0 296 54.4% 45.6% 0.0%
9 213 224 56 493 43.2% 45.4% 11.4%
9.5 184 207 0 391 47.1% 52.9% 0.0%
10 188 196 20 404 46.5% 48.5% 5.0%
10.5 149 162 0 311 47.9% 52.1% 0.0%
11 83 95 13 191 43.5% 49.7% 6.8%
11.5 31 32 0 63 49.2% 50.8% 0.0%
12 9 13 1 23 39.1% 56.5% 4.3%
12.5 3 0 0 3 100.0% 0.0% 0.0%
13 9 7 3 19 47.4% 36.8% 15.8%
13.5 7 3 0 10 70.0% 30.0% 0.0%
14 15 10 1 26 57.7% 38.5% 3.8%
14.5 8 14 0 22 36.4% 63.6% 0.0%
15 2 6 0 8 25.0% 75.0% 0.0%

[AussieVamp2]

so from that sample, it suggests you want to be getting lots to try this trick with a 9 involved, which makes sense - over 10% of these hit last year, and 13 was pretty ugly too, but small sample, same with 7 and 13s presumably were Colorado or Chicago pretty much

Do you need to look at when 9.5s land 10, etc.? Also depends which they take for their record keeping at covers.com I guess