<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Evan Lorad</title>
    <description>The latest articles on DEV Community by Evan Lorad (@evanlorad).</description>
    <link>https://dev.to/evanlorad</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4253607%2F1c969ef2-c0cf-4cac-b263-de7b9ad4aa28.png</url>
      <title>DEV Community: Evan Lorad</title>
      <link>https://dev.to/evanlorad</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/evanlorad"/>
    <language>en</language>
    <item>
      <title>We measured last GMP against listing gain on 1,264 Indian IPOs</title>
      <dc:creator>Evan Lorad</dc:creator>
      <pubDate>Thu, 19 Nov 2026 10:36:20 +0000</pubDate>
      <link>https://dev.to/evanlorad/we-measured-last-gmp-against-listing-gain-on-1264-indian-ipos-1cko</link>
      <guid>https://dev.to/evanlorad/we-measured-last-gmp-against-listing-gain-on-1264-indian-ipos-1cko</guid>
      <description>&lt;p&gt;&lt;em&gt;Evan Lorad, 18 November 2026. This is a cross-post; the canonical version is on Substack. All figures come from a public dataset, not from a backtest on private fills.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;Grey-market premium (GMP) is the unofficial rupee amount said to trade above an IPO’s issue price before the stock lists. It is &lt;strong&gt;not&lt;/strong&gt; an NSE or BSE price, and SEBI does not recognise grey-market trades. People still treat it as a forecast of the listing print.&lt;/p&gt;

&lt;p&gt;I wanted a boring question answered with a defined method: if you take the &lt;strong&gt;last GMP before listing&lt;/strong&gt;, convert it to a percentage of the issue price, and compare it with the actual listing gain, how often do the two point the same way?&lt;/p&gt;

&lt;p&gt;That comparison now covers &lt;strong&gt;1,264 Indian mainboard and SME IPOs&lt;/strong&gt; listed from 2 January 2023 to 11 November 2026, published on &lt;a href="https://ipogmpradar.com/" rel="noopener noreferrer"&gt;IPO GMP Radar&lt;/a&gt;. I operate the site. The full tables are on the &lt;a href="https://ipogmpradar.com/ipo-gmp-accuracy" rel="noopener noreferrer"&gt;accuracy page&lt;/a&gt;, last updated 18 November 2026.&lt;/p&gt;

&lt;h2&gt;
  
  
  Method, in one screen
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;&lt;code&gt;gmp_pct = last_gmp_rupees / issue_price&lt;/code&gt;&lt;/li&gt;
&lt;li&gt;&lt;code&gt;listing_gain = listing_price / issue_price - 1&lt;/code&gt;&lt;/li&gt;
&lt;li&gt;A &lt;strong&gt;hit&lt;/strong&gt; means the signs agree: both up, or GMP flat/negative and the listing flat/down.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Miss&lt;/strong&gt; = &lt;code&gt;listing_gain - gmp_pct&lt;/code&gt;, in percentage points.&lt;/li&gt;
&lt;li&gt;Hit-rate intervals on the sheet are 95% Wilson score intervals.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Coverage rules matter. An IPO with &lt;strong&gt;no&lt;/strong&gt; GMP quote is dropped; it is not imputed as zero. A genuine ₹0 quote is kept. Of the 1,264 rows, &lt;strong&gt;30&lt;/strong&gt; were captured live: every 30 minutes from 7 AM to midnight IST we read three public Indian GMP trackers and store the &lt;strong&gt;median&lt;/strong&gt;, with the IST timestamp on each mark. Subscription comes from the official NSE/BSE books. The other &lt;strong&gt;1,234&lt;/strong&gt; last-GMP values come from a public tracker’s archive, read 17 November 2026. We cannot independently confirm when each archive quote was taken, so a late revision cannot be ruled out.&lt;/p&gt;

&lt;p&gt;This is a historical match rate. It is not a trading system, and it does not predict the next listing.&lt;/p&gt;

&lt;h2&gt;
  
  
  Results
&lt;/h2&gt;

&lt;p&gt;The last GMP pointed the right way &lt;strong&gt;77%&lt;/strong&gt; of the time (95% CI 74–79%).&lt;/p&gt;

&lt;p&gt;The typical miss was &lt;strong&gt;7.9 points&lt;/strong&gt; (median; mean 13.9). Only &lt;strong&gt;37%&lt;/strong&gt; of listings landed within ±5 points of the GMP-implied gain (20% within ±2; 15% missed by more than 25 points).&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Board&lt;/th&gt;
&lt;th&gt;n&lt;/th&gt;
&lt;th&gt;Direction right (95% CI)&lt;/th&gt;
&lt;th&gt;Median miss&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Mainboard&lt;/td&gt;
&lt;td&gt;375&lt;/td&gt;
&lt;td&gt;83% (79–86)&lt;/td&gt;
&lt;td&gt;5.3 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;SME&lt;/td&gt;
&lt;td&gt;889&lt;/td&gt;
&lt;td&gt;74% (71–77)&lt;/td&gt;
&lt;td&gt;9.9 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The direction rate is not constant across the GMP distribution:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Last GMP band&lt;/th&gt;
&lt;th&gt;n&lt;/th&gt;
&lt;th&gt;Direction right (95% CI)&lt;/th&gt;
&lt;th&gt;Median miss&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;≥ 20%&lt;/td&gt;
&lt;td&gt;476&lt;/td&gt;
&lt;td&gt;97% (95–98)&lt;/td&gt;
&lt;td&gt;15.5 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;10–20%&lt;/td&gt;
&lt;td&gt;161&lt;/td&gt;
&lt;td&gt;84% (77–89)&lt;/td&gt;
&lt;td&gt;7.4 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;5–10%&lt;/td&gt;
&lt;td&gt;132&lt;/td&gt;
&lt;td&gt;70% (62–78)&lt;/td&gt;
&lt;td&gt;6.2 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;0–5%&lt;/td&gt;
&lt;td&gt;169&lt;/td&gt;
&lt;td&gt;50% (43–58)&lt;/td&gt;
&lt;td&gt;3.2 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Zero or negative&lt;/td&gt;
&lt;td&gt;326&lt;/td&gt;
&lt;td&gt;61% (55–66)&lt;/td&gt;
&lt;td&gt;5.1 pts&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Read that table twice. A last GMP of 20% or more was almost always the correct &lt;strong&gt;direction&lt;/strong&gt;, and the typical miss was still a wide 15.5 points. In the 0–5% band, direction accuracy was 50% — a coin flip with a tight confidence interval that still includes a toss-up.&lt;/p&gt;

&lt;h2&gt;
  
  
  A recent large-issue check, and a regime split
&lt;/h2&gt;

&lt;p&gt;Jio Platforms listed on 28 October 2026 at &lt;strong&gt;+21.36%&lt;/strong&gt; on a ₹1,119 issue price (₹1,358). The last GMP was &lt;strong&gt;+₹250 (+22.34%)&lt;/strong&gt;, stamped 9:37 AM IST that day. Miss: &lt;strong&gt;−1.0 point&lt;/strong&gt;. Direction: hit. One close row does not rescue the 0–5% band.&lt;/p&gt;

&lt;p&gt;Nearby, Maxwell Engineering Solutions listed +90.00% against a last GMP of +66.46%. Antala Industries had a ₹0 GMP and listed −20.00% (direction hit, 20-point miss). Vedant Poly Fabs had a last GMP of +1.06% and listed flat (miss, 1.1 points).&lt;/p&gt;

&lt;p&gt;The sheet also splits the sample by market state at listing. The Nifty 50 closed at &lt;strong&gt;25,387.4&lt;/strong&gt; on 17 November 2026, &lt;strong&gt;3.6%&lt;/strong&gt; below its 52-week high of 26,328.55 — a bull market on the site’s rule (within 10% of that high). In past bull conditions, direction was right &lt;strong&gt;79%&lt;/strong&gt; of the time (77–82%, n=1,063). In the 201 correction listings it was 63%. There is no bear-market subsample; the Nifty never sat 20% below its 52-week high in this window.&lt;/p&gt;

&lt;p&gt;The IPO tape is currently &lt;strong&gt;normal&lt;/strong&gt;: the last 30 listings averaged &lt;strong&gt;+10.62%&lt;/strong&gt; (hot is +30% or more, cold is under +10%). In past normal IPO markets, direction was right &lt;strong&gt;75%&lt;/strong&gt; (71–78%, n=571).&lt;/p&gt;

&lt;h2&gt;
  
  
  Listing day is not the holding-period result
&lt;/h2&gt;

&lt;p&gt;As of 17 November 2026, the &lt;strong&gt;median&lt;/strong&gt; name was +8.99% at the listing price, +10.88% at the listing-day close, and &lt;strong&gt;−1.94%&lt;/strong&gt; at the latest price. &lt;strong&gt;51%&lt;/strong&gt; of the 1,264 IPOs now trade below their issue price. Averages (+25.46% at listing, +46.04% now) are pulled up by a long right tail; the site treats the median as the fairer summary. Of the IPOs that listed up, 57% still stand above the issue price.&lt;/p&gt;

&lt;p&gt;For currently open or upcoming issues, the accuracy page shows past listing gains in the same GMP band on the same board — history, not a forecast. At 7:30 PM IST on 18 November 2026, Crystal Crop Protection (mainboard, open) had a GMP of +₹19 (+6.99%); 41 past mainboard peers in the 5–10% band had a median listing of +4.35%, a middle-80% range of 0.00% to +10.59%, and a 73% rate of listing up. Mahanadi Coalfields (upcoming) was +₹3 (+2.21%) against 71 peers in the 0–5% band (median +1.18%, −5.50% to +8.11%, 58% listed up). Watrana Rentals (SME, upcoming) was +₹37 (+15.10%) against 93 SME peers in the 10–20% band (median +11.73%, 0.00% to +36.88%, 82% listed up).&lt;/p&gt;

&lt;h2&gt;
  
  
  Takeaway
&lt;/h2&gt;

&lt;p&gt;The dataset supports a narrow claim: last GMP is a decent direction flag, especially when the quote is large or the issue is mainboard, and a weak point estimate of the listing print. It also does not tell you whether a listing gain will still be there months later.&lt;/p&gt;

&lt;p&gt;If you want the rest of the cuts — by year, by subscription multiple, by miss histogram — they are on the accuracy page linked above. Please treat every number as unofficial sentiment history.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Disclaimer.&lt;/strong&gt; Grey market premium is unofficial. It is not a price on the NSE or BSE, and SEBI does not recognise grey-market trades. IPO GMP Radar does not arrange grey-market deals. These figures are for reference only. They are not investment advice, not an invitation to trade outside the exchange, and not a promise of any listing price. Past accuracy does not predict the next listing.&lt;/p&gt;

</description>
      <category>analytics</category>
      <category>data</category>
      <category>python</category>
      <category>finance</category>
    </item>
  </channel>
</rss>
