Idea 1 feedback - Week 4

[KrissyDG]

Hey Zach, The Genius Gambler project seems interesting. I can see your point and intuition - Vegas certainly does rely on emotion and bad decisions. As smart data scientists, we should be able to win! I think your first paragraph lays that out really nicely. I might rearrange it just slightly to move the "Vegas loves offering bets to the uninformed & emotional public" sentence to line 2. Then you'd have: Vegas expects emotion/unwise decisions/arrogance They love that Most smart people have stayed away However, we have a better way! *Genius Gambler brings smart people/algorithms to betting. I think that flows really well through current situation -- problem -- solution. Regardless, I think your intro is strong. What you're saying matches my intuition, and even fires me up as a sort of pride thing. You've issued a challenge that smart people should be able to win bets, and of course I want to think I'm smart, so now you've got me convinced to join your project and take down Vegas! =)

I'm a bit lost in the middle section, though. I think I understand that you want to use machine learning to predict outcomes of games, and then build a separate algorithm to look for good odds. You started a description of the different algorithms, but I'm confused as to how they would play together. Machine learning uses a training set. Perhaps if you more clearly explained what training set you would use- what data it contained and where it would come from - it would be more clear. Then explain exactly what the inputs and outputs of your prediction model would be.

I think the end section mentions some great potential issues. "Wear and tear" and Play-offs would certainly be a challenge. There should be plenty of weeks of betting at the beginning of the season, though =)

Overall a neat idea. And professional sports certainly has TONS of data available. I think if you had more concrete tactical section, this could be a really cool project.

[sioharr]

This is a great idea for a project. Coming from Connecticut, I appreciate any proposal that says "if hedge fund guys can do it..."

I could picture a power-point graphic as part of your middle section that helps explain how the different algorithms work together along with the data inputs. I think this would help everyone understand what your vision is without giving away the secrets of your algorithms.

Based on this week's reading, I am interested in how many bets did Genius Gambler make last year to get that 60% return? Does your algorithm determine how much money to put down on every bet? Did Genius Gambler just have one really good bet and lots of bad little ones?

[zach-merritt]

Krissy and Siobhan, I really appreciate all the feedback. Krissy, I followed your advice on the sentence order in the first paragraph.

As a result of both your comments on my second paragraph, I completely restructured the second paragraph. I learned from both of your proposals and added a numbered bullet approach for clarity. I took a step back and thought more about the flow of my model. I tried to break the process out into three chunks and describe how they flowed together. I think I've added more clarity that should address both of your concerns.

Siobhan, I took your advice regarding my third paragraph and added additional detail regarding our success. I answered your proposed questions and more.

Thanks!

[scotthungerford]

Zach,

I appreciate the model of implementing informed and rational decision making to the world of gambling.

Gather the data:

It is interesting that this model will implement the the team-centric data, which is not typically considered when placing bets on sporting events. When you say privately sold data, what does that entail? Who supplies this data and what benefit does it bring to the algorithms? It is impressive that this algorithm will sweep for breaking news through twitter feeds and reputable sources.

Model team/player success: This Success Model seems to be capable of predicting the success of players and individuals. Once we gain a better understanding of data science principles. i'm sure I will be able to better understand how this works.

Identify bets and gamble Interesting to identify areas of undervalued opportunities. Does the algorithm consider how the recent gambling trends are going? Does the algorithm automatically place the bets or does it provide a probability of accuracy to the betters?

I