Insider: Back Testing the MTGO-to-Paper Ratio

Are you a Quiet Speculation member?

If not, now is a perfect time to join up! Our powerful tools, breaking-news analysis, and exclusive Discord channel will make sure you stay up to date and ahead of the curve.

Regular readers of this column will no doubt be familiar with the MTGO-to-Paper ratio, which is the price of a given set on MTGO divided by the paper price of the same set. This ratio is used to guide speculative purchases at key times of the year.

Although the ratio is imprecise, it gives insight into the MTGO market. I recommend reading up on it in this article where it was originally introduced, and in this more recent article where I have made some modifications to it.

The short version is that the ratio attempts to judge the value of a given set based on the price spread between paper and digital prices. For sets that are redeemable, MTGO prices should converge to paper prices over time as redeemers attempt to profit off of the spread. When the spread is large, as usually occurs in the Fall, then this signals a strong buying opportunity for speculators.

The Test

Today I’ll be judging the predictive powers of the metric by back-testing how it fared as a predictive tool in order to judge how sound the theory of redemption is.

In this article from November 16, I suggested that the MTGO-to-Paper ratio at the time pointed to Scars of Mirrodin (SOM) and Mirrodin Besieged (MBS) as the best speculative purchases to make. New Phyrexia (NPH) and Magic 2012 (M12) on the other hand were judged to have less value in comparison.

To see how the metric fared, I will compare theoretical purchase of the SOM and MBS mythic rare indices against the NPH and M12 mythic rare indices (Historical prices courtesy of MTGGoldfish.)

Index Nov 16 Price Post-Rotation Peak Date of Peak % Gain
M12 Mythics 3 5.5 May 10, 2013 83%
SOM Mythics 3.15 6.7 July 22, 2013 113%
MBS Mythics 3.2 6.5 Apr 4, 2013 103%
NPH Mythics 5.2 10.4 July 22, 2013 100%


Looking at the data, SOM has the edge in terms of gross gains, but both SOM and NPH are continuing to rise and probably will do so until the redemption cutoff date in November. The rise of the MBS mythic rare index was cut short by WoTC running out of redeemable sets. Once that happened, the prices of digital MBS cards began falling as they no longer had any value to redeemers.

On first glance, the ratio has held up as a predictive tool, but it's not a conclusive victory. With a medium-term outlook, speculating on any of the SOM block sets by buying the mythic index would have yielded nice profits. This suggests that my original recommendation to favor SOM and MBS over NPH wasn't entirely accurate, but clearly M12 mythics were the worst on a relative basis (though still profitable).

Rares Versus Mythics

A big part of the redemption story is that mythic rares are the bottle neck for redeemers. They are scarcer, and one needs all cards from a given set to redeem, so they should see the biggest prices moves.

With the available data, it's relatively simple to explore this idea further. Here's how the ratio did for the various indices, this time including regular rares as well as mythic rares.

Index Nov 16th Price Post-Rotation Peak Date of Peak % Gain
M12 0.8 1.4 May 12, 2013 75%
SOM 0.8 1.7 July 22, 2013 113%
MBS 0.8 1.7 March 28, 2013 113%
NPH 1.4 3.3 July 22, 2013 136%


Here we see a slightly different story. The gains for SOM are unchanged and the gains for MBS increase somewhat. M12 fares worse by including the regular rares, but NPH gets a substantial increase in returns.

In hindsight, the absolute best investment from the rotated sets would have been buying all the regular NPH rares, and avoiding the mythics! If you strip out the mythics from the NPH index, then the returns jump to 141%.

Again, the hypothesis is confirmed, but weakly. Focusing on mythic rares should on average increase returns relative to normal rares, but there can be notable exceptions.

Room for Improvement

So what's going on here? On the surface, it looks like the ratio is a mediocre predictive tool. But the sample size is small, and the results of back-testing the ratio are mostly skewed by NPH, an unusual set.

NPH has many powerful cards usable by a wide number of decks. Some of these are colourless, like the mythic rares Batterskull and Karn Liberated, but the powerful Phyrexian mana mechanic also means NPH has cards like Birthing Pod and Spellskite. Also, the 3rd set effect means that NPH is scarcer on average than either MBS or SOM.

A better test would have been to use the ratio to determine the sale date, and then compare returns. Also, returns should be annualized. If you could get 100% returns on SOM mythics in a few months, then theoretically you could take those gains and put them to work elsewhere, thus increasing your returns over the year. Comparing gross percentage gains across different time frames is somewhat problematic.

The ratio is evolving, and my understanding and use of it is changing too. Looking back on how it fared suggests that it has strongly confirmed the broad trends in the MTGO market, but that it is a weak tool in evaluating the merit of speculating on different sets.

The actionable plan for speculators should be to include rotating cards as a core holding of any portfolio during the winter months and to worry less about picking and choosing winners between sets.

Portfolio Update

This is a brief rundown of what I am buying, selling and watching in the market in the last week.


  • GTC Boosters are now the most expensive of the RTR block boosters, and I've completely sold down my stock on these. I've also sold down some copies of Thragtusk, most of my Geist of Saint Traft and Restoration Angel. Thragtusk has fallen far enough though that I think I'm better off holding my remaining copies for a price increase in late August, early September as players tire of the M14 drafting and look towards Standard for something new.



  • Rise of the Eldrazi (ROE) mythics and rares played in Modern. Prices have fallen on cards like Linvala, Keeper of Silence, but I expect strong rebounds by the time Modern season rolls around.
  • Innistrad block and Magic 2013 mythic rares. These will all be coming down in price, and I have been keeping track off the new digital-to-paper ratio for these sets on a weekly basis. Mirror-Mad Phantasm briefly touched 0.25 tix, which is a very strong buying opportunity. Most of the rest of the mythic rares do not look like good value at the moment, but if they hit 0.5 tix or less, they should basically be a snap-buy.
  • Prices on RTR and Dragon's Maze (DGM) boosters have started softening during the week of Cube draft. Expect this trend to accelerate once we get into the M14 release period. Scooping up a number of these with an eye to selling in September seems like a reasonable plan. Worst case scenario is I'll be left with some boosters that will appreciate in price over the winter.

10 thoughts on “Insider: Back Testing the MTGO-to-Paper Ratio

  1. Great article Matt! It’s always eye opening to compare numbers going a little bit beyond the charts. I love more and more whole set mythics and indices as a way to invest. I’m more and more convinced that it is an easy, predictable, and time saving way to make profits. I’m also, in part for my work, getting more interested in statistics and statistical significance, numbers, and studying variances and means. I might take a chance to run some stats between these different sets a see if there’s a real tendency somewhere of if it’s to random (and unpredictable)

    Average indices as presented by MtgoGoldfish are not “weighted” meaning they reflect more the variation of expensive cards than cheap cards.

    Still in this idea of comparing the average prices between Nov and post rotation peak, constituting a portfolio of each mythics bought for the same amount of total Tix would probably result in even highest profit that the one you mentioned, providing the rotated mythics behave the same way that M12 and M13 mythics when released (as I was saying in on of my posts: cheap mythics see higher increase, but are less significant in absolute price compare to expensive mythics).

    Also, and this is trickier to adjust, it is possible to yield bigger profits by selling individually cards when time is best for a given card. At least for M12, even if the average indice increases, some mythics (maybe because reprinted and/or standard playable) literally evolve at opposite of the indice (ex: Jace, Memory Adept)

    1. Thanks for the comment!

      Unfortunately we don’t have a ton of data yet, and so when looking at something like this, we are subject to bias due to the small sample size. But we do what we can, and the theory is lining up with reality pretty well on average.

      One of the hidden advantages to investing in an index of mythics is a reduction in transaction time. It’s much easier to buy/sell 4x of every mythic because bots are usually willing to buy/sell a bunch of different mythics rather than a lot of copies of one mythic. “Buying them all” reduces our own bias around pet cards, for better or for worse.

      Like, last year, I figured the SOM planewalkers would be good bets. But Mindslaver was probably the top performing SOM mythic, % gain wise. That was not what I expected.

      Lastly, for timing purpose, yes, individual cards do peak and trough at different times, especially ones that are not played competitively. In the summer of last year, the NPH praetors were cheap, and I loaded up on a lot of them. Vorinclex and Jin Gitiaxias both yielded 100%+ profit from August through November, an atypical time to be speccing on mythics in this way. By the middle of November rolled around, these ‘premature’ price jumps on the preators definitely reduced the potential gains on the NPH index as a whole from that point onward.

  2. Matt, lepong mentioned something that I think you missed.

    The index is an average of all cards of a certain rarity, and it’s going to be skewed / distorted by statistical outliers. it’s not ” weighted” in any way.

    If you bought one of every card of a certain rarity, you would replicate the index, but the thing is, no one ever does buy one of every card of a certain rarity, they buy more of the cheaper ones than the more expensive ones.

    Ands cheaper ones are going to be more affected by redemption, because changes in playability don’t affect them. Example: reap intellect. No one plays it, so out has hit absolute bottom.

    I think weighting the index should be its next evolution. let me know if you need help from someone who can use excel!

    1. I see! What do you think the weights could be? I guess price weights? Do you think there’s a practical advantage to adding weights? Generally I just use data that’s available and don’t collect much of my own. Adding weights would probably mean that I’d need individual data series for at least the mythic rares. Not sure hoe the cost/benefit of adding weights would work out.

      There is one example of a well known index that does not use weights, the Dow Jones Industrial Average.

  3. Weight the prices would tell you if it’s cheap or expensive cards that affect the average.

    Based on M12 and M13 mythics indices, cheap cards affected more the avrage than expensive cards, during their 1st 6-9 months (I didn’t look at it during post redemption period).

  4. Yes, I’m with lepongemagique, it would tell you whether you should be spending as much on cheaper mythics as tournament playable ones.

    Monsieur Sylvain wrote a long post in the forums, I guess you’ve seen it, where he concludes:

    “Second, and in addition to my reason 1- above there is a trend that I observed with all the M12 and M13 mythics: cheap mythics are more likely to yield better %…”

    I think some kind of weighted index would help to demonstrate this.

    But as to how one would weight it, I have no idea. I’m interested in statistics but don’t actually know anything about it… I think I’m going to have to have a go at it with excel…

  5. I’m not sure there’s a good way to look at “weighted” cards at any time for a long period. By changing their prices, cards would change their weight in the average making the whole thing probably difficult to interpret.

    But given a starting point you can track how cards, weighted or not (as their are from their origin) evolve and compare to the MtgoGoldFish.

    Probably not the easiest way to do it, but you could that you buy 1 tix of each of the 15 mythics of M14 at day 1 (whatever is that day).

    Then at any time point, look at the progression of each of the 15 mythics (+20%, +54%, -31%…) affect in consequence the 1 Tix starting value, add all the 15 values, and see whats the progression compare to the 15 Tix initially invested. And finally compare that progression to the M14_M indice progression.

  6. Good point… maybe the statistics should go on the premise that you’ll spend a certain amount (maybe 10 tix) of each mythic on a certain day in July, and compare how each ended up doing come redemption time. See if there’s any correlation between original price and final price.

  7. The index is skewed by statistical outliers like Karn etc, but this doesn’t really have an effect on redemption, since redemption requires one of each card.

    Somehow I got it into my head that junk mythics would be more profitable if bought after their relevant draft format revolved out and being sold to redeemers during the high season… I thought that they would be a safer bet because they’re not affected by changes in the meta… I’ve spent a bit of time on excel and found that isn’t really the case.

    Back to the drawing board.

Join the conversation

Want Prices?

Browse thousands of prices with the first and most comprehensive MTG Finance tool around.

Trader Tools lists both buylist and retail prices for every MTG card, going back a decade.

Quiet Speculation