Building The Otterline

[](https://substackcdn.com/image/fetch/$s_!H3Uc!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F212be991-cea9-4c01-945b-eb0fb165fe27_2050x1155.png)The first version of Otterline was created on October 15th and covered both the NFL and NHL. Since then, I’ve built five different versions of the system. Each iteration has improved the model significantly. Even today, Otterline is in a period of rapid expansion, with new algorithms, featured strategies, and filters continuously being implemented. The original idea was to create “Otterline AI” as a standalone prediction engine. However, once I began experimenting with consensus, the real value became clear—especially when factoring in the relationships between different models. Over the past two weeks, I’ve doubled down on this approach and repositioned “Otterline AI” as just one cog in a larger consensus-based system. This addresses a major issue with betting models: a high win rate alone is meaningless if you can’t narrow down the picks. Very few people, myself included, are betting every game, every night. A winning model only becomes useful once you can peel back the layers and provide confidence in a smaller, curated set of plays. That’s the goal behind our Diamond Picks and Power Plays. On a night with 13 games, the full slate is first filtered into Power Plays, which surface only the most confident picks based on consensus. From there, Diamond Picks further refine the list by combining the strongest consensus with the best value, highlighting the top bets of the night. I’ll continue monitoring the performance of these featured picks and make adjustments as needed. The reality is that systems like this perform best with consistent staking over time. Our objective is to stay above 50 percent, and we’ll continue working toward improving that edge daily. ### **Current Results** 3 Model Consensus: 172–124 (58.1%) 4 Model Consensus: 23–13 (63.9%) These numbers make me optimistic. The only thing we’re lacking now is more time and data. Overall, things are trending in the right direction as we tighten our focus and find our groove with consensus-based stats. As the data grows, we’ll be able to niche this down even further and truly lock in the picks. May the betting gods be on our side.