Machine-learned partial ticket value prediction

1. A computer-implemented method comprising:

accessing, for a stadium, a set of training data describing, for each of a plurality of historical sporting events at the stadium, attendance over a course of the historical sporting event and statistics associated with the historical sporting event;

training, using the set of training data, a machine-learned model configured to predict an expected likelihood that a seat within the stadium will be vacant over a course of a future sporting event based on real-time statistics associated with the future sporting event;

detecting, by a ticket exchange server in real-time, a seat within the stadium that was initially taken by a person during a sporting event but was then later vacated by the person during the sporting event, the vacated seat associated with a first ticket for the sporting event distributed to a first user and locked via a physical locking mechanism of the vacated seat, wherein the ticket exchange server is configured to lock and unlock each seat in the stadium;

providing, by the ticket exchange server, a rebate and refunding a deposit to the person in response to detecting that the person vacated the seat during the sporting event, wherein a value of the rebate is determined at least in part on an output of the machine-learned model; and

distributing, via the ticket exchange server, a second ticket for the vacated seat for a portion of the sporting event to a mobile device of a second user.

2. The computer-implemented method of claim 1 , wherein detecting the vacated seat within the stadium during the sporting event is in response to receiving an indication from the first user surrendering the vacated seat.

3. The computer-implemented method of claim 1 , further comprising:

in response to charging the second user for the second ticket, distributing the second ticket to the second user.

4. The computer-implemented method of claim 1 , further comprising:

accessing a previous value paid by the second user for a previous seat at the sporting event prior to detecting the vacant seat; and

charging the second user for the second ticket based on the previous value.

5. The computer-implemented method of claim 1 , further comprising:

determining, by applying the machine-learned model to real-time statistics of the sporting event, a value of the second ticket for the vacated seat before the first user has vacated the seat;

in response to determining that the value of the second ticket exceeds a threshold, sending, to a client device of the first user, an incentive to vacate the seat; and

in response to receiving an indication from the first user surrendering the vacated seat prior to a completion of the sporting event, offering, via the real-time ticket exchange server, the second ticket for the vacated seat for distribution.

6. The computer-implemented method of claim 5 , wherein the value of the second ticket is determined further based at least in part on a section of the stadium the vacated seat is in.

7. The computer-implemented method of claim 5 , wherein the value of the second ticket is greater if the sporting event has a below-threshold difference between scores of the sporting event than if the sporting event has an above-threshold difference between the scores.

8. The computer-implemented method of claim 1 , wherein the sporting event is one of a basketball game, a baseball game, a football game, a volleyball game, a soccer game, a tennis match, a hockey game, and a rugby game.

9. A non-transitory computer-readable storage medium comprising instructions that, when executed by a hardware processor, cause the processor to perform steps comprising:

accessing, for a stadium, a set of training data describing, for each of a plurality of historical sporting events at the stadium, attendance over a course of the historical sporting event and statistics associated with the historical sporting event;

training, using the set of training data, a machine-learned model configured to predict an expected likelihood that a seat within the stadium will be vacant over a course of a future sporting event based on real-time statistics associated with the future sporting event;

detecting, by a ticket exchange server in real-time, a seat within the stadium that was initially taken by a person during a sporting event but was then later vacated by the person during the sporting event, the vacated seat associated with a first ticket for the sporting event distributed to a first user and locked via a physical locking mechanism of the vacated seat, wherein the ticket exchange server is configured to lock and unlock each seat in the stadium;

providing, by the ticket exchange server, a rebate and refunding a deposit to the person in response to detecting that the person vacated the seat during the sporting event, wherein a value of the rebate is determined at least in part on an output of the machine-learned model; and

distributing, via the ticket exchange server, a second ticket for the vacated seat for a portion of the sporting event to a mobile device of a second user.

10. The non-transitory computer-readable storage medium of claim 9 , wherein detecting the vacated seat within the stadium during the sporting event is in response to receiving an indication from the first user surrendering the vacated seat.

11. The non-transitory computer-readable storage medium of claim 9 , the instructions further comprising instructions for:

in response to charging the second user for the second ticket, distributing the second ticket to the second user.

12. The non-transitory computer-readable storage medium of claim 9 , the instructions further comprising instructions for:

accessing a previous value paid by the second user for a previous seat at the sporting event prior to detecting the vacant seat; and

instructions for charging the second user for the second ticket based on the previous value.

13. The non-transitory computer-readable storage medium of claim 9 , the instructions further comprising instructions for:

determining, by applying the machine-learned model to real-time statistics of the sporting event, a value of the second ticket for the vacated seat before the first user has vacated the seat;

in response to determining that the value of the second ticket exceeds a threshold, sending, to a client device of the first user, an incentive to vacate the seat; and

in response to receiving an indication from the first user surrendering the vacated seat prior to a completion of the sporting event, offering, via the real-time ticket exchange server, the second ticket for the vacated seat for distribution.

14. The non-transitory computer-readable storage medium of claim 13 , wherein the value of the second ticket is determined further based at least in part on a section of the stadium the vacant seat is in.

15. The non-transitory computer-readable storage medium of claim 13 , wherein the value of the second ticket is greater if the sporting event has a below-threshold difference between scores of the sporting event than if the sporting event has an above-threshold difference between the scores.

16. The non-transitory computer-readable storage medium of claim 9 , wherein the sporting event is one of a basketball game, a baseball game, a football game, a volleyball game, a soccer game, a tennis match, a hockey game, and a rugby game.

17. A computer system comprising:

a computer processor; and

a non-transitory computer-readable storage medium storage instructions that when executed by the computer processor perform actions comprising:

accessing, fora stadium, a set of training data describing, for each of a plurality of historical sporting events at the stadium, attendance over a course of the historical sporting event and statistics associated with the historical sporting event;

training, using the set of training data, a machine-learned model configured to predict an expected likelihood that a seat within the stadium will be vacant over a course of a future sporting event based on real-time statistics associated with the future sporting event;

detecting, by a ticket exchange server in real-time, a seat within the stadium that was initially taken by a person during a sporting event but was then later vacated by the person during the sporting event, the vacated seat associated with a first ticket for the sporting event distributed to a first user and locked via a physical locking mechanism of the vacated seat, wherein the ticket exchange server is configured to lock and unlock each seat in the stadium;

providing, by the ticket exchange server, a rebate and refunding a deposit to the person in response to detecting that the person vacated the seat during the sporting event, wherein a value of the rebate is determined at least in part on an output of the machine-learned model; and

distributing, via the ticket exchange server, a second ticket for the vacated seat for a portion of the sporting event to a mobile device of a second user.

18. The computer system of claim 17 , wherein detecting the vacated seat within the stadium during the sporting event is in response to receiving an indication from the first user surrendering the vacated seat.

19. The computer system of claim 17 , wherein a value of the second ticket is determined based at least in part on a section of the stadium the vacated seat is in.

20. The computer system of claim 17 , wherein a value of the second ticket is greater if the sporting event has a below-threshold difference between scores of the sport event than if the sporting event has an above-threshold difference between the scores.