True or false?
Average Daily Theoretical (ADT) and Average Daily Worth (ADW) are reliable metrics for most of your database.
Answer: False
Across nearly every timeframe we’ve examined — whether it’s 90 days, 6 months, 12 months, or two full years — the same pattern emerges: Most casino players have low frequency. The majority have visited only once, twice, or maybe three times. In fact, nearly two-thirds of players fall into that range.
When we rely on ADT or ADW to assign value to those players, we’re building our decisions on shaky ground. We must always remember that short-term valuations are adversely affected by variables such as slot volatility, bet-to-bankroll ratio, unrated activity, and more. With so few data points, one off-day or a brief session can throw off the entire calculation. The result? Offers that miss the mark and marketing opportunities that quietly slip through the cracks.
True or false?
ADT stabilizes after just a few trips.
Answer: False
It might seem like a player’s value should even out quickly, but the data tells a different story. In our stabilization study, we tested how many trips it took for ADT to settle within a 10 percent variance. On average, it took eight trips to stabilize. When we narrowed the margin to five percent, the number jumped to 12 trips.
As we already know, most players never reach that many trips in a marketing period. That means ADT never actually stabilizes for most of your players.
This instability isn’t just theoretical. It’s something players feel. One month, they qualify for offers or an event, the next month, they don’t. Not because their behavior changed, but because our simplified math approach did. Ironically, most players are consistent in how they gamble. It’s the metrics we use that aren’t consistent — metrics that are as inherently random as the slot random number generator they’re based on.
True or false?
There’s a better way to value players than ADT.
Answer: True
We believed there had to be a better approach, and we built one. Using machine learning, we created a model designed to predict player worth earlier and more accurately than traditional methods. We utilize metrics based on how customers play, not just on the outcomes. We find that player gaming behavior is much more consistent than the outcome of those sessions, and we can leverage that predictively.
Our model learns from players with long, stable play histories and identifies early behaviors that correlate with long-term value. These include average bet, coin-in per minute, session length, game selection, cash buy-in, bet to bankroll ratio, free play usage, travel distance, age — nearly any metric we can collect. Once trained, it applies that knowledge to new or low-frequency players and forecasts their future potential.
It’s not guesswork. It’s grounded in data, and it’s amazingly accurate.
True or false?
Our model outperforms ADT and ADW at every stage.
Answer: True.
We ran head-to-head comparisons using a statistical measure called R², which tells us how accurately each method predicts future value. We studied the accuracy of each valuation method as trip frequency increases, starting at just one trip. The closer the R² is to one, the better the prediction.
Here’s what we found:
- ADT starts weak and doesn’t exceed 0.90 R² until after trip 12
- ADW is better in the early stages, but still lags behind
- Our machine learning model hit 91 percent accuracy by trip three
- After just one trip, our model achieved an R² of 0.74 — nearly double ADT’s 0.43 at the same point, and well above ADW’s 0.37.
That means you can act with confidence earlier and give consistent offers that reflect player behavior, not just volatile math.
True or false?
Most casinos still rely on ADT for marketing and tiering.
Answer: Sadly, true
Despite better tools and more data than ever, ADT remains the default in many rating systems. It’s familiar, it’s entrenched, but it’s also flawed, especially when used for early decision-making. And ironically, that’s exactly when smarter valuation matters most.
The consequence? Some players receive inflated offers based on a lucky loss, while others with real long-term potential are ignored. Machine learning helps reverse that, offering a truer reflection of player worth from the start.
True or false?
Machine learning is only for data scientists, not marketers.
Answer: False
Here’s the exciting part: You don’t need a technical background to take advantage of machine learning. While the algorithms behind the scenes are complex, the results are straightforward and actionable. For marketers, this means more accurate and consistent player lists updated automatically with revised predictions that often outperform traditional metrics like ADT or ADW. It’s still an invite or segmentation list, but now it’s smarter. Some players will rise, others will drop, but the overall quality and consistency of your targeted marketing will improve dramatically.
True or false?
Predictive models can improve marketing campaigns and grow the bottom line.
Answer: True
With a better understanding of early player value, you can refine invite lists, personalize offers, reduce unnecessary comping, and drive more meaningful engagement. That means less wasted budget and more profitable trips, because the offers align with what players are likely to do, not just what they have already done.
It also creates a more stable experience for players. When their behavior remains consistent, so do their offers.
True or false?
We’re just getting started.
Answer: Absolutely true
This model is just the beginning. We also built virtual slot machines so we can run large-scale slot simulation models to understand how volatility, house advantage, pricing, and free play impact player value and time on device. By modeling variables like wager rates and bet-to-bankroll ratios, we’re uncovering new insights into player valuations and gaming experiences that will drive entirely new marketing and segmentation strategies.
The future of player valuation isn’t just based on averages. It’s based on patterns, probabilities, and a deeper understanding of human play behavior.
Turning data into dollars
Whether you’re a marketer, analyst, or executive, embracing advanced machine learning techniques sets you apart from the status quo. It unlocks new perspectives on value, loyalty, and long-term growth — perspectives that drive real financial impact. It’s about identifying the next VIP in your database — the high-value customer you haven’t even noticed yet.
Raving’s Data Analytics team understands casinos. With 40+ combined years on the floor, we understand the gaming customer and how to evaluate, track, and measure your marketing strategy to increase the bottom line.
With changing player behaviors and expectations, there’s no time like the present to measure, evaluate, and implement change at your business. With seamless integration, our team can begin quickly, delivering consistent, monthly reports and recommendations. Let’s talk! Contact Liz Palar today at liz@betravingknows.com for more information or visit our page here!
