Like low level chess, the brazilian presidential election will be defined by who blunders less

The two main candidates for the presidency of Brazil are Lula and Flávio. Under 1 month ago Flávio was on the lead based on online betting platforms, now Lula pulls ahead but the chart shows it is going down. All thise up and down is due to the fact that in turns each one of them says or does something that makes them lose support.

Until the elections, in October, both of them can blunder enough so that the odds will be vastly different from what they are now. Not even taking into account the possible rise of competitors for the "third place".

I know that using online betting platforms is not exactly scientific, but I believe that with (low) millions of trading volume there must be bots taking into account real popularity research and real statistics. I am assuming the market is self correcting and accounts for more variables than a single opinion poll/research, but of course in reality markets are not efficient, it is just the easy way to gauge their odds of winning.