Can only a human understand a human?
by Chris Naudi
Recently, at ICE VOX, I was part of a panel discussing whether A.I. will be able to replace a sportsbook Trader and an extremely interesting debate ensued. Essentially, we were a group of 3 ‘old school’ sportsbook industry veterans challenging an ex-Harvard University student, now Head of Data Science at an up-and-coming A.I.-driven sportsbook Service provider, on whether A.I. would be responsible for mass redundancies in trading rooms. An interesting hour-long discussion followed!
It would be fair to say that, while us Trading ‘oldies’ were all slightly sceptical of the full potential of A.I. at the beginning of the debate, we were all receptive to the fact that we believe that the sportsbetting environment is drastically changing and A.I. could potentially provide the key to unlocking more in terms of product offer and pricing accuracy. During the discussion, the A.I. expert excellently debunked some of the myths and misconceptions that we brought up and, with a generally ‘pro-A.I.’ audience in attendance, the topic was tackled from different angles.
The main counter-arguments that we made during the hour were as follows:
Only a human can understand a human!
A slight sweeping statement here, and hardly backed up by the fact that we all agreed that we don’t think even a human could understand certain logic when it comes to our ‘punting’ brains. The point we were trying to get across is that if it is difficult enough for a human to try to understand our behaviour/logic, then how would a machine or an algorithm be able to do so? While, yes, patterns can certainly be found in both recreational and non-recreational sportsbook behaviour, you certainly cannot segment all punters and a level of open-mindedness needs to prevail.
A.I. effectively has no understanding of the problem that it is trying to solve and it will therefore produce a result that cannot be explained in human terms (Black Box). A.I. is only looking for patterns of behaviour and displaying the results of those patterns.
Relevance of data analysed
Another question to ponder is at what point does data that you would feed your A.I. algorithms become ‘stale’ or less relevant. Again, you could counter-argue that the same rule of thumb applies to the algorithms that are more traditionally used in pricing models and compiling.
The punter’s emotional state
Can A.I. ever understand the emotional state and real feelings of the punter watching the game in the pub or at the ground? With a roaring crowd, changing weather conditions, psyched up players and real raw emotion at a game spurring players on – it’s at this point that a punter sees or feels something that cannot really be quantified, and triggers an emotion to want to place a bet. This, for me, is where a trader will always need to step in and take over from what AI can deliver.
Anyway, there is no doubt in my mind that the potential that manual tagging and A.I. can unlock in the offer of data-driven markets to an extent we have not previously seen in the betting industry. Betting markets such as ‘The hardest shot of the match’, ‘The distance covered on the pitch by an individual player’, or ‘The longest successful pass’ can all be produced in near-to real time with video footage of a Football match.
What we all agree on is that the role of the trader will evolve in the coming years as A.I. becomes more prevalent in the trading room. Will it mean mass redundancies in the trading room? Unlikely, however the role of the trader will probably be more to support the power of A.I. So what do you think?… Can only a human understand a human?
Check out our latest videos below:
Related vacancies:Productdiscover Fraud & Paymentsdiscover Sales, Business Dev & AMdiscover Legal & Compliancediscover Marketing & PR, Communicationdiscover Cafe Manager | SOHOdiscoverTechnical Payment Analyst | Trustly Group ABdiscoverSales Manager | IkajodiscoverSouth America Business Development (Remote) | iTech labsdiscoverCorporate Lawyer | InPlayNetdiscover