Custom software & automation case study
Gambletron is a sports research and decision-support platform that brings statistical analysis, sportsbook market comparison, recommended picks, bet tracking, and performance reporting into one dashboard.
Project facts
The problem
MLB games are played nearly every day, and researching a full slate properly means pulling information from a lot of places: starting pitchers, advanced statistics, bullpen usage, confirmed lineups, weather and ballpark conditions, umpire tendencies, sportsbook prices, and market movement.
The client was spending two to four hours on this every single day. The research platforms supporting that workflow cost over $1,000 per month combined, gated their most useful information behind additional premium tiers, and still did not consistently publish every recommendation. Even after paying for several services, the research stayed fragmented across websites, spreadsheets, and sportsbook screens.
The information itself was publicly available. The hard part was collecting, normalizing, and analyzing it consistently for every game, every day, without missing anything.
The build
From engine to platform
BlueSail's first deliverable automated the client's MLB research methodology as a scheduled analysis engine. Each day, the system collected the slate's baseball and market data, evaluated every matchup using a deterministic scoring framework built from the client's methodology, published qualifying recommendations, recorded the results, and graded itself after games completed.
The client liked the results enough to change the goal. Instead of a personal tool, he wanted a product other bettors could use. BlueSail built the full Gambletron platform around the original engine: daily matchup research, sportsbook comparison, expected value, arbitrage, line movement, bet sizing, bet logging, grading, bankroll reporting, and closing-line-value analysis.
The proprietary scoring model remains the engine underneath, but it is kept physically separate from the product. The dashboard makes the research understandable and actionable without exposing Gambletron's formulas or weights.
How it works
The system collects and normalizes data from multiple baseball, statistical, weather, and market sources: MLB schedules and starting pitchers, advanced pitcher statistics, recent velocity and Statcast performance, team offensive and defensive performance, bullpen usage and condition, confirmed lineups and handedness, weather and park factors, umpire tendencies, sportsbook prices, line movement, and final scores.
Environmental analysis goes deeper than a temperature check. The system carries venue metadata for every MLB stadium: location, center-field orientation, roof status, and park factors, so wind speed and direction are interpreted relative to how each ballpark actually sits.
No language model decides which team should win. The system applies the same established research framework, consistently, across the entire slate: the analysis a disciplined human researcher would do, executed without fatigue or shortcuts.
Designed to abstain
Gambletron assesses every game, but it does not manufacture a daily pick. Moneyline recommendations must score at least 90. Run-line recommendations require at least 95. Totals are evaluated internally but are not published as official recommendations because their results have not yet met Gambletron's publication standard.
If no game qualifies, the system publishes nothing that day. Even on a slate full of qualifying games, official recommendations are capped at five. The system is built to be selective, not prolific.
The confidence score is a ranking produced by Gambletron's proprietary framework, not a calibrated probability or a guarantee of any outcome.
Daily workflow
Development results
Over a three-month development period, the model was tracked through three iterative phases covering 229+ MLB games, finishing that period at a 64.3% win rate on published recommendations.
Selection criteria tightened progressively across those phases. Earlier versions used a lower qualifying score, while the current system publishes only recommendations that meet stricter confidence standards.
These figures describe a completed development period, not current or future performance. Gambletron does not place wagers or make decisions on behalf of its users.
The engineering challenge
The hard part was never the dashboard. It was making independently changing data sources behave like one reliable system. Team and player identities had to be reconciled across providers that do not agree on names or IDs. Data arrived at different times, in different formats. Lineups could be incomplete early in the day. Announced starters could change hours before first pitch.
Optional sources could fail outright. Sportsbook API usage had to stay within quota every day, indefinitely, handled through tiered caching and automatic credential rotation when a provider throttles a request.
BlueSail designed the system to degrade safely: preserve previously published recommendations, prevent post-start rewriting, and keep operating when nonessential data is unavailable. A missing weather feed should not take down the day's slate, and it does not.
What it replaced
In the client's words
We hired BlueSail AI to help us turn a pretty manual sports research process into an actual software platform.
Before working with Elias, we were pulling information from a bunch of different sources, comparing odds and stats, running our own calculations, and then trying to piece everything together. We had a methodology that worked for us, but turning that into actual software was a much bigger project.
One of the things that impressed me most was how quickly Elias understood what we were trying to accomplish. He was able to take our research process and turn it into actual algorithms, scoring logic, automated workflows, and analytics.
This definitely wasn't just a website with a few APIs connected to it. He built the systems that collect and organize the data, the custom logic behind our models, and the analytics we use to actually evaluate the results.
The platform now handles things like matchup scoring, expected value calculations, no-vig odds, sportsbook comparisons, line movement, bankroll and performance analytics, automated grading, historical results, and a lot of the research that we used to do manually.
There are also automated processes running behind the scenes that keep pulling in and checking new information, which means we aren't spending all day going through different websites and updating things ourselves.
We originally started the project around MLB, but it has grown quite a bit since then. We're now working with NFL and NCAA football as well, and the system was built in a way where we can keep adding new sports, data sources, analytics, and models over time.
The project also ended up being more complicated technically than we probably realized at the beginning. Different data providers use different formats, information doesn't always update at the same time, and there are a lot of random edge cases that come up. Elias was good at working through those issues and building checks around the data instead of just assuming everything would work perfectly.
From a business standpoint, the platform has replaced several tools we were paying for (over $1,000/mo) and saves us hours of manual research. More importantly, we now have our own system that is built around the way we actually work instead of trying to force our process into a bunch of different software products.
I think Elias's biggest strength is his ability to take a complicated idea and figure out how to actually build it. He understands AI and automation, but he also has a strong understanding of algorithms, data, analytics, APIs, and custom software development.
I would recommend BlueSail AI to anyone that has a complicated process, a lot of data, or a proprietary way of doing something and needs someone who can understand it and turn it into a real working product.
Current status
Gambletron has been fully handed off to the client, who now owns and operates it at gambletron.ai. At handoff the MLB model was live and had been beta-tested with approximately 20 users, and the platform was built to expand to all major sports.
The dashboard's odds-comparison and research tools already supported football, basketball, hockey, and college sports at handoff, with scoring models for those leagues in development.
BlueSail builds custom software and automation around workflows that generic products cannot handle.
Gambletron is a research and decision-support tool. It does not place wagers, guarantee outcomes, or provide financial advice. Past development results do not guarantee future performance. Gambling involves risk, and users should never wager more than they can afford to lose.
