more-arw search

Q&A Forum

What statistical forecasting tools, models & techniques would you suggest to forecast secured cash flow

Hewitt Hsiao's Profile

I work in an online gambling company. In this industry, paying money to your customers in time is the most important thing, so our company needs to reserved lots of cash to run daily business transaction. But now we want to precisely forecast daily cash flow in our account based on historical data, so that we can utilize money to do other investments. Our business model is that each member has to deposit money to their account first and then they can bet different games from their account. Once they want to withdraw money from their accounts, we have to make sure we have enough cash to pay it.

Now the questions are what statistical tools, models and techniques you would recommend we use to forecast our daily cash flow.

I can only think we can use our daily cash transaction data to compute normal distribution and standard deviations for the past three years.


Topic Expert
Wayne Spivak
Title: President & CFO
LinkedIn Profile
(President & CFO, |

I don't particularly disagree with you potential model, but:

1. You will never "precisely" forecast anything.
2. Your model will not forecast outlier transactions. In other words, winners who shouldn't have won at the "tables" but did break the house advantage.

An example is a Superball a few years ago. I consequently was in Las Vegas and so I couldn't resist some silly bets on the underdog. All the "smart" money went for the other team, so my odds were greatly increased and felt I could blow a $100 for "fun".

My team won and the Casino was caught off-guard and it took 45 mins for them to bring up from the vaults enough cash to pay us off.

Good luck!

Topic Expert
Jake Feldman
Title: Managing Director
Company: Global TaxFin Advisory Group LLC
(Managing Director, Global TaxFin Advisory Group LLC) |

Hi Hewitt,

If I understand this correctly, this sounds more like a cash management inventory stocking issue that depends on a probabilistic distribution forecast of demand for cash. There are sophisticated cash inventory models that boil down to choosing a lower bound where you must replenish the cash to some midpoint level and an upper bound where you invest surplus cash to reach back to the midpoint level. The probabilistic distribution may be around some exponential smoothing or more sophisticated auto-regressive trend line function that considers seasonality and trend factors. As Wayne indicates, because you may have some extreme cash outflows in the probability distribution, the stockout avoidance factor will dictate holding more cash.

Do some research on cash management and then contact me if interested.



Get Free Membership

By signing up, you will receive emails from Proformative regarding Proformative programs, events, community news and activity. You can withdraw your consent at any time. Contact Us.

Business Exchange

Browse the Business Exchange to find information, resources and peer reviews to help you select the right solution for your business.

Learn more

Contribute to Community

If you’re interested in learning more about contributing to your Proformative community, we have many ways for you to get involved. Please email to learn more about becoming a speaker or contributing to the blogs/Q&A Forum.