[r2d2]

Does anyone know if pomeroy, on their website, has predicted score results vs. actual results? For example, if at the beginning of the season pomery predicted unc to beat duke at home 78-74, and the actual was 90-68? curious to know in other words based off his predicted scoring model how well he is hitting ATS on a % basis? any help is appreciated.

[Axis]

I've often wondered that myself, but I think that is a stat you'd have to keep on your own...Which would be quite time consuming, but I haven't found anything for that. Maybe I just haven't searched hard enough.

[kroyrunner89]

yeah he doesn't have something tracking how well his predictions did, after the game he just replaces the prediction with the actual score. his predictions are usually really close to the line vegas sets anyways, so you won't find a ton of value there.

[r2d2]

yeah he doesn't have something tracking how well his predictions did, after the game he just replaces the prediction with the actual score. his predictions are usually really close to the line vegas sets anyways, so you won't find a ton of value there.

i agree with his margin of victories being close to the line. however, i think the value would lie in his "probability of victory" %. In other words, just to the right of his predicted scores, you will see a % chance to win the game number. So, you could take say anything with a 70% chance or better of winning outright and track that comparison ATS. Just a thought i had.

[curious]

i agree with his margin of victories being close to the line. however, i think the value would lie in his "probability of victory" %. In other words, just to the right of his predicted scores, you will see a % chance to win the game number. So, you could take say anything with a 70% chance or better of winning outright and track that comparison ATS. Just a thought i had.

You don't need to do that. You can take any line and convert it into a probability of winning. The linesmakers already tell you what they think the probability of winning is.

A 70% win probability is a line of -234. If you blindly took all favs -234 or higher you would get massacred.

There is no value in taking pomeroy's win-loss score predictions by themselves blindly. You have to actually cap the game. That is what I do. I start with his pythag number but I adjust it for several factors. I look for lines that are off. The first half of the season there were a lot of lines that were off by a large margin. Lately the lines have been very tight. Non conference games are a gold mine, for some reason the linesmakers don't cap them correctly. Not sure why. Conference play it gets harder and harder to find value as the season progresses.

Right now big upsets are happening at a more than normal % clip. I mean complete scrub teams are blowing out solid teams. However, the Pomeroy stats are not predicting these upsets. At the beginning of the year the pomeroy stats were predicting the upsets. For example, a few days ago Indiana State beat Illinois State, and not just by a basket or two, I mean they BEAT them. Pomeroy's stats showed Illinois State as the play, the stats were not even close. In terms of pythag, strength of schedule, win %, predicted line vs actual line, it wasn't even close. Yet Indiana State blew Illinois State off the court. This happens of course, my point is that you used to be able to predict such upsets using pomeroy, that has become much harder to do.

Bowling Green just lost at home to a scrub team. Well, the list is pretty extensive lately. At the beginning of conference play big favs, if you filtered them using pomeroy's stats, almost never lost SU. I think at one point my money line favs > -200 were something like 73-3. At the beginning of the year, during non-conference play it was easy to find big money line dogs that had no business being a dog at all, let alone a big dog. These cashed at > 50% on the money line. ATS was an even higher percentage. Pomeroy's stats would tell you that a particular team should be a favorite and the line would have that team as a 9 point dog (Maryland a few times). Bingo, you just found a moneyline dog to bet on. That isn't happening much anymore.

What I found about following pomeroy is that the lines makers go through cycles in how religiously they stick to the pythag and log5 formulas. The reason that pomeroy's predictions are very close to the linesmakers' lines is that that is how the linesmakers calculate the lines. They vary from this, yes, because of their bias. At the beginning of the year big money line dogs were easy to find and cash because "popular" teams were being grossly overpriced. Then big favs became easy to find and cash because dogs became grossly under priced. Now there seems to be parity.

yes, you can find value every day but you have to use pomeroy as a starting place and then cap the game to figure out where the value is.

Here is an example. On the 17th, Pomeroy showed Purdue and Michigan State as being too close to call. He had Purdue as a small fav. But pythag wise and strength of schedule wise, Michigan State is the stronger team. So, pomeroy's numbers didn't tell you anything.

[curious]

Does anyone know if pomeroy, on their website, has predicted score results vs. actual results? For example, if at the beginning of the season pomery predicted unc to beat duke at home 78-74, and the actual was 90-68? curious to know in other words based off his predicted scoring model how well he is hitting ATS on a % basis? any help is appreciated.

That isn't the stat you want. The stat that you want is, how often did Pomeroy's prediction show that there was value in the linesmakers' lines and how often did that prediction turn out to be true.

For example, on the 17th Pomeroy showed that there was huge value in the Murray State line. According to Pomeroy the linesmaker's were off by 600 points (money line). Murray State easily covered that game. SEMO has been covering big spreads so just capping the game you would not have seen the value. I was afraid enough of SEMO I even bought 3 points on this game, turned out I didn't need it Murray State blew them out as Pomeroy predicted.

The reason that I say you don't want to know how often his scores are correct is that what is important is how his scores relates to the lines being offered and do his spreads show that there is value in the lines being offered, and is he right. You need to know if his line showed that there was value in the spread being offered by the linesmakers.

Until about two or three weeks ago I could state very strongly that Pomeroy's lines showed you plays that had value every day. Starting about three weeks ago it is getting harder and harder to find value using his lines alone.

There have been a lot of upsets lately, i'm talking big upsets, where Pomeroy did not predict the upset. If you had been on the fav on the money line in those games you would have been killed. (Ask me how I know this. LOL)

Until about 3 weeks ago you could bet all of Pomeroy's big favs on the money line knowing you were safe because his stats pretty much filtered out the upsets. You were usually on the dog in the upset because Pomeroy predicted it.

You can't do that any more. The lines have tightened up CONSIDERABLY in the past few weeks and Pomeroy is not predicting big upsets very often. So, you can't play big favs on the money line based on Pomeroy alone or you will get killed.