I am a clinical psychologist with an interest in human behavior especially as it relates to evaluating effective treatments for individuals with cognitive and neuropsychiatric difficulties. For the past 12 years I have been conducting clinical research using functional magnetic resonance imaging (fMRI) to explore brain-behavior relationships, in particular the response of the brain to various psychopharmacological treatments for individuals with various brain disorders. To do this, I employ sophisticated statistical tests using specialized statistical analysis programs to analyze signal in various parts of the brain (mainly AFNI and SPSS)...
Motivation and Approach:
About 10 years ago, when I started designing experiments and examining the relationships between signal in different parts of the brain, two things became clear: 1) Although we can never account for all of the variables that predict human behavior/brain response, it is nevertheless still possible to account for a good number of them and predict it with some level of accuracy by taking into account as many key variables as possible; 2) How these variables interact with one another to be able to predict outcomes/responses; 3) essential ways to characterize these variables and the data that they express in order to optimally highlight differences between groups/conditions when they exist.
Brain/behavior to Sports prediction:
I realized that one could apply these very same techniques/research methods and principles to predicting the outcome of sports events as long as one could account for enough of the key variables that can effect their outcomes. While I have aspirations to analyze sports data in the same traditional way as I do with brain research (when I find the time), right now, this exceptional service provided by Killersports.com allows me to do the next best thing, namely to write models that account for various factors that are important in predicting the outcomes of these sporting events frequently enough to yield a significant profit. Even though no real statistics are offered and an index of the power of sample size are not provided using this medium, we can still arrive at some strong findings (through win percentage and ROI) that point towards one team or another or to the over/under of totals of games.
I have been seriously applying my research skills to sports prediction over the past 3-4 years. It has been a challenging and fun hobby, and I am serious about making a profit from my selections that are heavily influenced by my query results. I have learned over time how to effectively separate this work from my emotions, as I feel that my emotions related to the outcome of games very often introduce unnecessary bias to my selection process.
To this end, I have created hundreds of these effective models/queries and continue to create them on a daily basis when I have time. I routinely incorporate variables like streak, goals, penalty goals, minutes, assists, shots on goal previous match-up results, day, month, season, site streak, previous wins and losses and others to help me to successfully predict MLB and NHLgames.
These queries have been helping me to make a profit in sports wagering. I make sure that all of my queries have been successful year in and year out, have at least a 65% win rate and have at least an 8% ROI, while always maximizing sample size. Occasionally I will showcase a high win percentage query that has a relatively smaller sample size but with a high win rate--in these instances, the low sample size is due to the limited appearance of the particular situation expressed by the query using the available data-set (This often is a function of the fact that MLB database data only go as far back as 2004).
Needless to say, I am always refining these queries and coming up with new ones in order to increase profit as well as learning from other experienced guys like JMon and Cherry (I have found that a humble approach and taking into account other perspectives as from these guys has only served as a good checks and balances function for what is an imperfect yet profitable science).
I hope you will find them as useful as I have.
In order for you to maximize profits/minimize losses, it will be critical that, besides the results of my queries, you take into account additional information available on the particular day of an active query, including things like: 1) any significant injuries; 2) the track record of the pitchers; 3) results of previous same-season matchups; and 4) time-specific variables including but not limited to month, day and 5) closeness of current line to that of the average historical line.