Who pays the vig !

[sportshobby]

I wonder if his testing of previous seasons didn't assume each play could be put in and evaluated before determing the amount of the next (as opposed to a full day's worth at a time).

If he has an average of 3-4 plays per day, there must be some days when six plays come. Now we have 96% of our bankroll in action in a single day's set of plays. If he is your friend you'd better have a talk with him.

[SLAM DUNK]

Reno, 16% on 52.6% plays even at an average of +117 is ridiculously high. Just for comparison purposes, I use 14% of my bankroll on 11/10 game I rate at 80%.(even though everyone claims there is no such thing). I came up with my percentages running random trials, an looking for the worst possible outcome. I like to assume the worst. That way no surpises. Another problem with your approach is that you are making equal plays, on games which are not equal. In other words I can assure you that his +130 games win at a higher clip than his +170 games. By combining them to determine the optimal percentage, you are overplaying the +170 games and underplaying the +130 games. Also it sounds like their is nothing predictive about the program. In other words are all games 52.6% or are some 54.5 % games and others 49.5%. My advice, do some more refining, and bring that BR % DOWN. With that many plays it will not take long to build up the profits, and also remember the PAST does not always EQUAL the FUTURE.

[reno]

sportshobby, the program was based on all plays each day being put in at once. He is well aware that true Kelly is based on single plays. But again, he is simply following what his computer results showed produced optimal return over the 4-year period in which there were over 2000 plays.

[wintermute]

reno

Your friend is doing it exactly right. I admire his cojones. I tend to be very conservative with my wagers mainly because I usually have way more than 3 or 4 wagers in a day and because I have lots of parlays. If you have a potential return on parlays of 10 percent the optimal Kelly is about 3 percent of capital, not 10 percent.

I too have a database of wagers over the past 3 or 4 years and have come to the same conclusion he has. Kelly always gives the best return.

It doesn't matter very much that the expectation varies from bet to bet. As I mentioned earlier in this topic, there's not much difference between a) wagering your average expectation for every bet and b) varying your wager based on the individual game expectations.

This of course assumes that you don't bet your entire capital at once. A good rule would be to never wager more than 50 percent of your capital. If you have 5 wagers and your optimal wager is 16 percent of capital then you would wager at most 10 percent of your capital per game.

It doesn't matter that you go 0 for 4 in a day. In fact that's bound to happen fairly often. What matters is the long term. Take the following example. Say you win half your wagers at odds of +132 - this gives you a 1.16 expectation - same as your friend. By the way, I'd like to know your friend

Over 100 games you'll win 50 times. If you bet 0.16 percent of your capital per game your capital after the 100 games will be a factor of

( ( 1 + 0.16 * 1.32 ) ** 50 ) * ( ( 1 - 0.16 ) ** 50 )

your original capital. This works out to 2.370049 times original capital. The answer is the same no matter what the order you win or lose games.

The only thing that can kill you assuming that you can in fact maintain a 1.16 expectation is that you lose so badly over the short term that you can't scrounge up the money for the sportsbook's minimum bet. You are out of the game if you have no capital.

'mute

[wintermute]

reno

I just reread your post. Your buddy's system if run over all 2000 games would increase hib original bankroll by a factor of 2.370049 ** 20 or 31271464.

Bill Gates, eat your heart out.

'mute

[wintermute]

To be fair, the comments I made earlier about the order of wins and losses not mattering is true only if you're wagering one game at a time.

If you're wagering for example 4 games per day, it would be much better for you to win 2 games and lose 2 games on each of 2 successive days than it would be to win 4 games on the first day followed by 4 losses on the second day.

The return over the long haul is still great but wagering multiple games per day will lead to much more violent swings in the value of your capital.

'mute

[AussieVamp2]

'mute, you win that much the only person you would be able to bet with is Bill Gates, and I think he prefers cards

[buckeye]

Zippy,
I understand your reasoning, you could even "think" of it as $5 being taken from the loser's bet and $5 being taken from the winner's payoff. Miller's focus is on who is affected, i.e. has different results, depending on the size of the vig? The winner! Therefor he is asserting that the party affected by the size of the vig is "paying" it. I can't say that "thinking" about portions of the loser's bet as being vig changes this argument. Maybe I am just dense!

[zippy]

buckeye:

Yeah, we’re just going back and forth on it. I haven’t seen a reason to change my position on the matter yet, but then again irish insists that his example “proved” the Miller position. So it’s appearing unlikely that either side’s going to get through to the other.

You’re not dense. You’re just stating what makes sense to you, same as I am.

Most gamblers seem to intuitively regard vig as a penalty that the loser pays, which of course I think is just the same mistake that Miller is making on the other side.

Maybe it all comes down to what you regard as the relevant non-vig bet to compare it to. The Miller proponents look at a wager where the player risks $110 and the book risks $100, and they see the non-vig version of that being a wager where both sides risk $110. This makes it appear that vig alters what a player collects when he wins, but has no effect when he loses. But it’s just as accurate to say that the non-vig version of that wager would be one where both sides risk $100, which would make it appear instead as if the vig costs the player $10 when he loses and has no effect when he wins.

The size of the vig is irrelevant. If it was 12/10 instead of 11/10, then you could either say that you are being shorted even more on your wins, or you could say that you’re paying an even greater surcharge on your losses, and both would be equally correct.

One indication to me that it is a trivial semantic difference and no actual substantive difference is that it seemingly changes nothing whichever way you describe it. For those who are convinced that Miller is right that the winner pays the vig, what do you do differently as a result of that realization than you would do if you were clueless like me and still thought it makes equal sense to say the loser is the one disadvantaged by the vig? Would you bet more games, fewer games, higher amounts, lower amounts? Would you have to win a different percentage of your plays to break even?

If a sports book announced “We’ve decided that from now on, instead of having the winner pay vig as is traditionally done, we’re going to have the loser pay the vig," what precisely would change?

Anyway, if I’m wrong, I’m wrong. I won’t waste space with further posts on the matter unless someone comes up with something new. In any case, buckeye, if the winner truly does pay the vig, may you have a long and successful career paying vig on bet after bet after bet.

[rabbit]

Regarding the 16% of BR, I think mathematically it might work out on a long season like baseball, where the program is churning out a 52% win rate on dogs. However, I think it might be a tough go psychologically. Since we're dealing with subjective probabilities instead of loaded dice, a bettor may lose faith in his or her program after a couple of 0-4 days, thinking that perhaps something fundamental in the game has changed (e.g., expansion, band-=box stadia, etc.). Thus the bettor may cut back on his or her plays at the most inopportune time. I know 16% wouldn't be right for me.

I'm sure some of you old-timers used to read Huey Mahl (RIP). He used to go on-and-on about the Kelly method in his tout sheet, how to adjust for simultaneously plays, etc. However, he seemed to assume that it was easy to pick 60% winners, and, though he had a quantitative bent, wrote little about objective ways to actually pick winners.

Did Huey end up drinking himself to death?

[wintermute]

Two books open up in your town.

One book - book A has a special for the month of October - bet with no vig on any NFL games you want.

The other book - book B has the normal line - 110 to win 100.

Joe hears about book A's special and bets 10 games at $100 a crack and wins all of them. At the end of the month he's wagered $1000 and is up $1000. Larry is not so lucky, he bets 10 games too but loses them all. He's down $1000 at the end of the month.

Buck and Pete don't know about book A. They are customers of book B. Buck wagers 10 games at $110 a crack and like Joe wins all of them. He has wagered $1100 to win $1000. Pete like Larry loses his 10 bets. He wagers $1100 and loses $1100.

Aha you say, losers pay the vig because Pete lost an extra $100 and Buck won just as much as Joe. Not true. Pete lost an extra $100 because the vig forced him to put down $1100 instead of $1000. Buck lost $100 because the vig lowered his potential profit from $1100 to $1000.

Everyone pays vig and it's the same on both ends.

'mute

PS My last comment on this. It's apparent from previous posts on this topic that once people take a side on this issue they don't change their minds.

It'd be interesting to see if there is any consensus. I'll start a new thread where people can vote for what they believe. Vote W for winner pays, L for loser pays, B for both pay.

[irish]

Zippy I went to this guy Millers website,alot of interesting stuff there,sound like he is or has worked with mathematicians.If you know who infact pays vig (brokerage fee) it is the starting point in your money management plan.I am an advocate of the flatbet plan (with sports that is ,horses are a different kettle of fish)so if I assumed the loser was paying fee my long term plan would be knocked out of wack,I would be betting $1 to either win $1 (false) or $1.10 to lose/chance of winning $1, when in reality with the flat bet plan I am actually risking $1 to win 91c and with this system any win percentage over 53% I am in the black.Hey I may be thick headed as an irishman but I am not stupid,,,,,,,,,,Just kidding Sincerely Irish,,,,,

[rfd]

Pete lost an extra $100 because the vig forced him to put down $1100 instead of $1000. Buck lost $100 because the vig lowered his potential profit from $1100 to $1000.

If Pete truly lost $100 and Buck lost $100 that's $200 worth of vig on $2200 worth of bets, which is not the case.

I'll side with Irish. If only because my wife is Irish and I haven't won an argunent in 20 years.

rfd

[reno]

rabbit, I told Huey to stick with one kind of booze, but he insisted on doing round-robin parlays with the sauce. Seriously, I heard that he was an alcohlic who ruined his health with the stuff. And I also heard that he didn't have a whole lotta money at the end.

[sportshobby]

Maybe he kept on betting 16% of his bankroll of each play . . . )