A journalist asked me, "how do we know your forecasts/predictions/projections are based on data analysis and not luck." It's a fair question, so here are a few receipts in no particular order.
In 2021, I correctly called both Senate races in Georgia...
In October, I correctly called the Senate race in PA including the margin.
Brazil as well...
There is more if you use the search feature. I am sincerely tired of being asked to show my work when my "peers" are just taken at their word while consistently getting it wrong. Winning the lotto is "luck;" analyzing tons of data has nothing to do with luck.
Frank got everything wrong, and he gets to go on as if nothing happened. I correctly predicted every Senate race except Ohio, and I am being asked to show my work.
And for the record, my model had Ohio leaning Republican, but I gave it to Democrats because of Dobbs.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Now Democrats need 2 more seats to reach 218. Yes, it gets complicated from here, but it's not impossible.
I believe CA-41 can still flip to Democrats, but the recent updates have been meh. Redistricting makes predicting what will happen here difficult, but I give it to Dems...
With CA-41, Democrats have 217 seats and need one more seat. I'm afraid I have to disagree with others, and I believe when more votes are counted, Democrats will win AZ-6, and that gets Democrats to 218. But let's assume I am wrong about CA-41 or AZ-6...
I find it amusing and somewhat telling how the experts who claimed a red wave was coming, and Republicans would gain 30+ seats in the House and 4+ seats in the Senate without providing evidence, now want to lecture others about "accuracy." Come on, really?
These are the same experts who also got 2016, 2018, and 2020 wrong. They knew many of the recent 2022 polls were junk and cited them anyway. And if they claim they did know the polls were junk, then we have a bigger problem because it was clear the polls were junk...
Either way, they are in no position to criticize anyone about anything related to elections.
As for my House forecast, I will tweet a detailed thread later this morning about how I arrived at 219.
Everyone keeps asking me about the House. Before Election Day, I predicted Democrats would hold the House with 2 - 6 seats. My prediction was based on Democrats performing terribly in Florida, doing reasonably well in NY, and securing the win in CA. The problem is...
Democrats did worse than I expected in NY, and now they must flip a few seats like R-CO 3rd district and overperform in CA to pull it off. Right now, the odds favor Republicans, but it's not over. There is a realistic path for Democrats to win, but it will be tough...
If Democrats win, now it will be with 1 - 2 seats. If Republicans win, the max seats are 4. If I were Republicans, I wouldn't be celebrating yet.
Something is happening on the ground that contradicts the polls. Democrats are actually gaining steam in some states. If Democrats have 2018-level turnout on Election Day, it's going to be a good night for Democrats. Republicans have a lot of ground to make up on Election Day...
Republicans can erase Democrats' lead on Election Day; it happens often. The problem for Republicans is the influx of newly registered voters. If a substantial number of these voters wait to vote on Election Day and vote Democrat, then Republicans will have a bad night...
Moreover, waiting for your voters to vote en masse on Election Day is a huge gamble. If there are long lines and/or bad weather, many voters may not vote. Based on the early vote data, it is safe to assume turnout will be high in many states on Tuesday...
.@elonmusk charging $20 a month is way too high, but I do agree a paid option for certain features is the way to go. In my opinion, a one-time fee to get verified is acceptable, and maybe a nominal yearly renewal fee. $49 to get verified and $9.99 a year to renew...
But I sincerely believe a Twitter App Store is how you can generate significant revenue. I had to build an entirely new browser extension and app for Bot Sentinel just so my users could use it with Twitter. It would've been significantly more straightforward...
If there were an SDK that would allow developers to extend the functionality of Twitter and sell their apps, that would be a new revenue stream for Twitter. More importantly, developers would improve the security of Twitter, and it wouldn't cost you anything...
NateTheGrifter and his election-denying lawyer stated I never released software that had millions of users. They paid someone for a deep dive and still got the basics wrong. Let me help you out...
I released Nexus Radio in 2008, and it remained one of CNET's top streaming audio software for over 5 years. A year after I released it, it won CNET's Webware award. Nexus Radio was eventually acquired by an investment group.
CNET wasn't the only website Nexus Radio was popular on.
By the time I released Bot Sentinel, I had spent 8 years prior developing encryption algorithms and wanted to do something fun. I literally coded Nexus Radio for fun.