<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Matt Sly on Medium]]></title>
        <description><![CDATA[Stories by Matt Sly on Medium]]></description>
        <link>https://medium.com/@mattsly?source=rss-c9afc23988fa------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*OT_cJ-X3pBFeRcXn.jpeg</url>
            <title>Stories by Matt Sly on Medium</title>
            <link>https://medium.com/@mattsly?source=rss-c9afc23988fa------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 11:51:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mattsly/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Introducing The Parent Bank!]]></title>
            <link>https://medium.com/@mattsly/introducing-the-parent-bank-750d9bf18e39?source=rss-c9afc23988fa------2</link>
            <guid isPermaLink="false">https://medium.com/p/750d9bf18e39</guid>
            <category><![CDATA[personal-finance]]></category>
            <category><![CDATA[parenting]]></category>
            <category><![CDATA[piggy-banks]]></category>
            <dc:creator><![CDATA[Matt Sly]]></dc:creator>
            <pubDate>Fri, 20 Jan 2023 20:44:02 GMT</pubDate>
            <atom:updated>2023-01-26T18:17:33.294Z</atom:updated>
            <content:encoded><![CDATA[<p><em>I wanted to teach my kids about personal finance and I decided the old school piggy bank is inadequate. So I decided to build </em><a href="https://www.theparentbank.com/"><em>The Parent Bank</em></a><em>!</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/180/1*yWN86sn7M1L57w_fY6uKVQ.png" /><figcaption>The Parent Bank Logo!</figcaption></figure><p>OK, so kitsch and nostalgia aside, a Piggy Bank is pretty lame when you think about it.</p><p>Sure we want our kids to save money. <strong>But we also want to teach our kids that money should earn more money. </strong>A piggy bank doesn’t do that. At all. It’s basically akin to teaching our kids to keep their money under a mattress.</p><p>(Also — do we really think our kids will pay for goods with cash and coins will be something they use as grown ups?)</p><p>We could sign our kids up for savings accounts with banks, but those accounts decide on their own interest rates, which are pretty low and unimpressive for small amounts of savings like our kids have.</p><p>The Parent Bank is super duper simple. You are the bank so you can make the rules!</p><p>The most important rule is setting up your own a <strong>savings rate </strong>that is appropriate for you and your family. My kids, they earned 100% interest (yes you read that right!) until they had saved $100. I set the rate this high because I really wanted to demonstrate to them that interest is indeed “Magic Money!” (what we call it in the product).</p><p>Once they hit that $100 threshold, I dialed it down to 12% per year, which is still high enough that for each spending decision, they are able to start thinking about about how much “future money” they are taking from their future selves if they want to withdraw and buy something…just like all of us do.</p><p>You can also <strong>automate allowance</strong> if you want (I pay their age every 2 weeks). And then you can share a secret link with your kids to monitor their account and sign up for periodic email updates about their account.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ygho4zUAQyfy1hCqmiO-fw.png" /><figcaption>What the kids see!</figcaption></figure><p>For older kids, you can easily disperse The Parent Bank savings to Apple Cash for spending. So there is no need to sign up for and pay for one of those kid’s debit cards and give them another thing to lose. Here’s how that works:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/1*RtVapGttVZpqd-Fe8jk_vg.gif" /><figcaption>Send to Apple Cash for spending!</figcaption></figure><p>Of course eventually my kids will be old enough that they will need a ‘real” bank and a credit card — I’m thinking probably 16 or so? But between the ages of about 8 and that future point in time, this is what we are going to use in the Sly Family.</p><p>So if you have kids (or not) I hope you will check out <a href="https://www.theparentbank.com/">The Parent Bank</a> and let me know what you think (even if it’s why you will never use it!) 🙏</p><p>Will this end up having millions of users like my last side hustle <a href="https://www.futureme.org/">FutureMe</a>? No idea…probably not…but The Parent Bank has been fun to build and has kept me busy for a few weeks and hopefully some other parents out will there find it useful!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=750d9bf18e39" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[“Salary Adjusted” Baseball Standings]]></title>
            <link>https://medium.com/@mattsly/salary-adjusted-baseball-standings-bb4a2a6913bb?source=rss-c9afc23988fa------2</link>
            <guid isPermaLink="false">https://medium.com/p/bb4a2a6913bb</guid>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[baseball]]></category>
            <category><![CDATA[moneyball]]></category>
            <category><![CDATA[oakland-athletics]]></category>
            <dc:creator><![CDATA[Matt Sly]]></dc:creator>
            <pubDate>Mon, 16 Oct 2017 14:39:06 GMT</pubDate>
            <atom:updated>2018-01-08T15:15:57.015Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Leveling the playing field. So to speak.</em></p><p><em>(Note: Originally written in 2013)</em></p><p>In baseball, as with most things in life, having money is a considerable advantage. Teams with more money tend to win more, as well they should. This post provides an overview on a methodology for a theoretical application of a “salary handicap” to adjust baseball standings based on how much money a team pays out in salary.</p><p>There is a wide distribution in the payroll of various teams. The 2012 MLB season salary distribution is shown below. That year, the lowest paying team (The Oakland A’s) paid a total of $49,137,500, which was almost 4x less than the $195,998,004 team salary from the highest paying team (The New York Yankees).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/773/1*jBGS6bERsek-XQUs_ly6CQ.png" /><figcaption>Data source: <a href="https://www.usatoday.com/sports/mlb/salaries/">USA Today Salaries</a></figcaption></figure><p>Considering data on both Salary and Winning Percentage over time, we can see that there is indeed a correlation between how much money a team pays out and how many games a team wins. The following chart shows the results from three teams over the 15 year span between 1997 and 2012.</p><p>The big spending New York Yankees have won a lot of games as we would expect. The low spending Pittsburgh Pirates have not won much. And the Oakland A’s, of <em>Moneyball</em> fame, despite not spending much, have still managed to win a fair number of games.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tq7-2-fHEiZiQkC8Wscxxw.png" /><figcaption>Note that to normalize for different years, the independent variable on the x-axis is salary as a percentage of the league mean for that season, rather than an absolute number. The y-axis is simply the winning percentage of a team in a given year.</figcaption></figure><p>Using a simple best-fit natural log regression (a good choice since it will asymptote, and even an infinite salary level will not produce more wins than games played…just don’t tell the Yankees…) we can determine a formula to predict how many games a team will win.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/849/1*AHGY0c8DDT7_KN-XAcKEsg.png" /></figure><p><em>Expected Winning Percentage = .082 * ln(Team Salary as a Percentage of the League Mean) + .5065</em></p><p>Then, using a team’s expected winning percentage, we can “handicap” that team’s current record, using the following ratio:</p><p><em>(Actual Winning pct./Expected Winning pct.) = (Adjusted Winning pct./.500)</em></p><p>In other words, instead of assuming that the team would be performing at an average level if it had a .500 record, we assume that it would be performing at an average level if it has exactly its expected winning percentage given its salary. Teams that are exceeding their expected winning percentages will have adjusted winning percentages greater than .500, and likewise, teams that have an actual winning percentage less than their expected winning percentage, which have an adjusted winning pct. that is less than .500.<br><br>For example, the 2003 Yankees. True to form, their team salary of $151 million dollars far exceeded that of any other team. And by plugging this salary into the above formula, their expected winning percentage was .645. Their final records was 101–61, or .623. Though an impressive record in absolute terms, relative to their expected winning percentage, the Yankees actually performed worse that expected, and their Adjusting Winning Percentage was just .482 (.500 * .623/.645 — see above ratio) good enough for a third place finish in the AL East behind Boston and Baltimore.</p><p>Applying this methodology to the final standings of the 2013 Season, we see that the “Salary Adjusted” Playoffs would have looked a lot different than the actual playoffs — the Indians (not the Tigers) win the AL Central, with the Royals getting the Wild Card.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1BaapxqNq-UzUX1_Djk2Zw.png" /><figcaption>Salary Adjusted AL Central Standings for the 2013 Season</figcaption></figure><p>If you are interested in exploring further, this Google sheet that I put together has all the data, an interactive bubble chart of all of the data points, and live adjusted standings.</p><p><a href="https://docs.google.com/spreadsheets/d/e/2PACX-1vREwBXt6Dc5F1vkV_mDIzinEehEaFK2SDvbpG8v1zlL7ilmut9ZAF_FjaDcpjBBPAGu5lHRt7A8nH2x/pubhtml?gid=578475944&amp;amp;single=true&amp;amp;widget=true&amp;amp;headers=false">Major League Salary &amp; Winning Percentages</a></p><p>Disclaimer: I grew up in Oakland, and am a lifelong A’s fan :)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bb4a2a6913bb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Product Usage vs. Product Utility]]></title>
            <link>https://medium.com/@mattsly/usage-vs-utility-1d1a7aa65330?source=rss-c9afc23988fa------2</link>
            <guid isPermaLink="false">https://medium.com/p/1d1a7aa65330</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-design]]></category>
            <dc:creator><![CDATA[Matt Sly]]></dc:creator>
            <pubDate>Sat, 03 Jun 2017 12:44:15 GMT</pubDate>
            <atom:updated>2026-01-21T22:35:54.883Z</atom:updated>
            <content:encoded><![CDATA[<p><em>It sounds simple, but don’t mistake product “engagement” for product quality</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/457/1*as5zH1X1aFYa3KB62OVbkg.png" /></figure><p>A common metric cited to measure software product success is Monthly Average Usage — “MAU” — how many users does a product have in a given month?</p><p>It’s a quasi-standard metric, but I challenge teams to think hard about the difference between <em>Utility</em> and <em>Usage</em> — they aren’t necessarily the same. And in fact, I’d argue over the long run, can be inversely correlated.</p><p>My favorite example to cite is TurboTax. I have serious respect (and equal parts pity) for the team that has been able to build out, maintain and update the state machine that emulates all the permutations of the state and federal tax codes.</p><p>TurboTax is a great product. I give it an enthusiastic 10/10. But I really, really, really don’t want to use it very often or for very long. In fact, <em>the </em><strong><em>less</em></strong><em> I use TurboTax, the </em><strong><em>better</em></strong><em> I feel about the product</em>. As TurboTax gets better and better, usage should go down, as it’s increasingly easier to do my taxes. In fact, the ideal would be that TurboTax just does my taxes for me and I never have to even use the product at all.</p><p>If you looked at TurboTax monthly usage between April (after the tax deadline) and May, I’m sure it would take a nosedive. A lazily designed analytics system would set off red alerts and meetings would ensure and PMs would get in trouble.</p><p>Another product example are the many (too many, IMHO) CRM ecosystem products. Ultimately, you want your customer teams talking to and helping customers — not using software. Sure, software is a useful tool to help them be more productive in doing so…but it’s a means, not an end.</p><p>And, getting more provocative here, even with social networks — does increased usage necessarily correlate with utility? If I have to spend more time wading through updates or ads I don’t care about, maybe my usage goes up, but my utility goes down. If Facebook is able to tune its newsfeed algorithm so effectively that I can spend less time in the product, I’m a happy customer.</p><p>Our fundamental goal as product developers is to allow people to be productive and enjoy their life. Software is a means to end. Indeed it is “eating the world”…but, ironically perhaps, in the long run, the more pervasive software becomes, the less “usage” (in the traditional sense) occurs — it’s everywhere, but, in the long game, should really feel like it’s nowhere. As Arthur C. Clarke penned: ”Any sufficiently advanced technology is indistinguishable from magic”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1d1a7aa65330" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Price of Spam]]></title>
            <link>https://medium.com/@mattsly/the-price-of-spam-f9e51bd58a99?source=rss-c9afc23988fa------2</link>
            <guid isPermaLink="false">https://medium.com/p/f9e51bd58a99</guid>
            <category><![CDATA[spam]]></category>
            <dc:creator><![CDATA[Matt Sly]]></dc:creator>
            <pubDate>Mon, 21 Apr 2014 01:20:39 GMT</pubDate>
            <atom:updated>2017-10-15T17:39:53.207Z</atom:updated>
            <content:encoded><![CDATA[<h4>Would You Pay $1 a Year for Guaranteed Spam-free Email?</h4><p>A few months ago, my Microsoft colleague, <a href="http://datamining.typepad.com/data_mining/2010/03/spam-email-has-a-0000008-conversion-rate.html">Matthew Hurst</a> shared a link to an <a href="http://www.newscientist.com/data/images/archive/2749/27491501.jpg">infographic </a>visualizing data from a spam-bot called “Storm.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SBzsXX6K3uh7rHNz40Mh-Q.jpeg" /><figcaption>source: New Scientist: <a href="https://www.newscientist.com/data/images/archive/2749/27491501.jpg">https://www.newscientist.com/data/images/archive/2749/27491501.jpg</a></figcaption></figure><p>The noted sample set contained 35 million spam e-mails, which was 1.5% of the total number of messages the Storm sent out that month. Extrapolated, this a total of <strong>28 billion messages per year (35 x 12 / .015) </strong>sent form the Storm spambot.</p><p>The graphic also notes that “Storm” led to <strong>$ 3.5 M in sales </strong>for that year, which implies a .000008% conversion. So dividing annual sales by annual emails ($3.5 M / 28 billion e-mails), we can estimate that the <strong>marginal revenue per email of $0.000125</strong>. Or $1.25 for every 10,000 messages.</p><p>This math implies that if the cost of sending spam was to rise above $.000125 per message, Storm would cease to be a profitable operation. But one of the main challenges with spam is that it costs spammers essentially nothing to send out. No matter how effective anti-spam software is, as long as a spam campaign leads to one single sale, the marginal profit is, well, infinite.</p><p>So, technical details aside, what if we could a require “e-mail postage” as a means to cut down on, or even end, spam? I can feel the backlash from the Libertarian-leaning freemium crowd. What?! Tax?! The Internet needs to be free!? What about the unencumbered exchange of ideas amongst people!?</p><p>But in fact the cost impact would be minimal. If a typical user sends 25 personal email messages per day (about what I send), multiplied by 365 days and by the marginal cost required to make Storm unprofitable ($.000125), it would cost…$1.14 per year in “spam tax.”</p><p>What do you think — worth it?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f9e51bd58a99" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Shoot the two!]]></title>
            <link>https://medium.com/@mattsly/shoot-the-two-ed4aaa553294?source=rss-c9afc23988fa------2</link>
            <guid isPermaLink="false">https://medium.com/p/ed4aaa553294</guid>
            <category><![CDATA[basketball]]></category>
            <category><![CDATA[probability]]></category>
            <category><![CDATA[analytics]]></category>
            <category><![CDATA[statistics]]></category>
            <dc:creator><![CDATA[Matt Sly]]></dc:creator>
            <pubDate>Mon, 21 Apr 2014 01:08:56 GMT</pubDate>
            <atom:updated>2024-10-22T19:21:10.458Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/desat/multiply/yellow/60/overlay/yellow/1*oJJC9Yi5mhUL1oanXmHO7w.jpeg" /></figure><h4>Game Theory and “Winners Outs” Strategy for Half-Court Basketball</h4><p>I used to be good at basketball.*</p><p>I still love to play, but now in my old(er) age, I’m more of a “cagey veteran.” Which means I do things like applying statistical analysis to my shooting strategy.</p><p>A topic I’ve been thinking about recently: how does “winners outs” (when you make a basket you get the ball back) that is the common rule in a half court game affect the optimal shot taking strategy? Scoring <em>and earning another chance to score </em>is a different outcome than scoring and having the other team get the ball (like in a regular full-court game).</p><p>To explore the difference, let’s consider the example configuration:</p><ul><li>You make 50% off your 2 point shot attempts</li><li>You shoot 33% from beyond the arc on 3 point shots</li></ul><p>In addition, for the sake of simplicity, let’s assume:</p><ul><li>There is no difference in the probability of getting an offensive rebound based on what kind of shot you take</li><li>The probability of a turnover is the same regardless of what kind of shot you are planning to take</li></ul><p>In a full court game, with this configuration, you are entirely indifferent as to whether to take a 2 or a 3. You expect exactly 1 point per shot: 50% × 2 points = 33% × 3 points = 1.</p><blockquote><em>Note: a decent counterpoint is that your “utility” of making a shot is not just the points you earn, but also the “glory”, and that making a 3 yields more, and thus even in this configuration, you’d shoot the 3 because of the potential “glory utility”…but we’ll leave that out of scope for now!</em></blockquote><p>But in a half-court game, with “winner’s outs”, this calculation changes. If you “shoot the 2&quot;, you have a 50% chance of not only of scoring 2 points, but also getting the ball back, which means another chance to score (and another, and another…as long as you keep making shots) So your expected number of shot attempts per possession is a <a href="http://en.wikipedia.org/wiki/Geometric_series">geometric series</a>: 1 + .5 + (.5)^2 + (.5)^3 … — which means that you expect to have 2 possessions (1 / ( 1 — r) where r = .5), and thus score <strong>2</strong> points each time you take a 2 point shot in a winner’s outs game.</p><p>By contrast, if you “shoot the 3&quot;, you only expect 1.5 shots (1 / ( 1 — r) where r = .33). You still average 1 point per shot attempt, but that means only 1.5 expected points per possession following a strategy where you only shoot the 3.</p><p>This is huge! A 25% difference in the expected points per possession in the half-court game based on the same shot taking strategy that results in a wash for a full-court game.</p><p>But wait…we’re not done yet! Since we’re playing up until a point total (common half court games are played to 21) rather than a timed game, you need to consider how many more points you need to win. If you have 18 points, a single 3 can win the game (33% chance), whereas you’d need two 2-point baskets. (50% × 50% = 25% chance to win during this possession). So if you have 18 points, your best strategy is to shoot the 3!</p><p>OK, let’s keep this analysis going. In many games, 3s count as 2, and 2s count as 1. My gut has always been that this is unfair, and that when playing these kinds of games, I always shoot a “2” from behind the arc. But is that the right strategy? Let’s do the same Math with the same configuration: Shooting the 3: 2 pts * .33 pct * 1 / (1 — .33) shots = 1 expected point per possession. Shooting the 2: 1 pt * .5 pct * 1 / (1 — .5) shots = 1 expected point per possession. The same expected outcome! So in fact, whomever came up with the rule that 3s should be 2s and 2s should be 1s in a half-court game…is kind of a genius!</p><p>How does the other team’s expected shooting ability factor in? And what if I’m feeling hot from behind the arc today and shooting 40% from the 3? This all deserves some more analysis…and even <a href="https://github.com/mattsly/expectiminimax.js">some coding</a>(!)…but we’ll stop this post here for now…and go outside to shoot some hoops!</p><p>* <a href="https://lh3.googleusercontent.com/pw/AM-JKLX2s2td5bhe7v3vkY3bXnJv3oiai392Ta9eRGBtTwf5t3F2wbT7u95d4FTYFLv_YpeaUbjwuJ3nkj47zHNY8d0vhvEIzFiB6B-8k1yo5axK7B3-fRWYzofbvGuya0NCvxQMMmitfvip7B_OYKc5FYGQeg=w586-h751-no">proof</a> that I used to be good a basketball (swear that’s me!)…and that my high school team unfortunately never got the memo about long shorts being the in thing…</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ed4aaa553294" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>