Variable stars are very common. One could even say that every star is variable, if only our instruments were sensitive enough to register the smallest possible changes in magnitude. In principle, this means we can point to almost any non-moving object in the sky and say “it's variable!”, if only we are able to confirm variability at some level.
Using modern photometry, we have already found dozens of millions of variable stars. New catalogs are continuously being released, and even more will be released soon, when next instruments will be able to monitor very slight changes previously unreachable (like Kepler "microvariables"), or around fainter stars, most notably through the LSST at the Vera Rubin Observatory. While it is still relatively easy to keep track of comets, and increasingly difficult for asteroids... but we lost the battle for variable stars a long time ago. There are simply too many of them.
The Variable Star Index (VSX) is an excellent tool that aims to be as complete as possible, and the MGAB Variable Star Catalog is attached to it. However, because of the massive amount of photometric data produced nowadays, large observatories and collaborations no longer publish each variable star individually. There is no time for that. Instead, variables are published in bulk catalogs. They are more time efficient, but cause a conflict I encountered early on. Only a small fraction, 1% or even less, are variables truly worth extensive individual analysis and the rest could be put into one bag. In result, the first really large catalogs are often the least complete, especially in terms of variability type - at least for me, a very important indicator when you're submitting new discoveries, particularly when there are too many objects to analyze within limited time and one must begin selecting targets based on personal interest. Once objects are analyzed automatically, nobody returns to them—unless someone else does so with their own data during cross-matching.
The sheer quantity of detected variables has also led to another consequence: discovery credit is rarely treated as important. This is the central subject of this page. When discovery credit is not tracked at an early stage, the problem cascades. Others follow the same approach, and credit gradually disappears into the fog. VSX does an excellent job in this regard, yet large collaborations rarely treat it as the primary source for discovery vetting. There's probably no other better site for this purpose. Instead, papers often cross-match only a few of the largest catalogs, many of which never tracked discovery credit in the first place. For astronomers, this is additional, time-consuming work that has little impact on their own research output, so it is often ignored.
This seems like a problem of the past - today we have better data, tools, also more things are now automated. A DSCT pulsating variable can be easily distinguished from a reflection (R-type) variable around HW Virginis binaries in many ways - something that was far from clear in the earliest large catalogs. The MGAB Variable Star Catalog was established and expanded in years, when it was really a problem:
Large catalogs containing hundreds of thousands to several million new variables were published with very basic information - less than amateurs were providing manually at those times.
Many early catalogs contained low-quality, contradictory, or even absurd variability data (it is still a discovery of variability).
Catalogs clearing almost the whole sky with (especially high-amplitude) variables appeared before we had modern data tools, relying on old datasets such as Gaia DR1 or even UCAC4.
Discovery credentials were often ignored entirely (papers mattered more, because it's easier to add a citation and "it looks better").
To illustrate the situation, consider an amateur astronomer using average astrophotography equipment during the years when most MGAB discoveries were made. In my case, it was an 8-inch Newtonian telescope with a cooled ZWO ASI1600MM camera. With this setup, I was able to detect variability in a large number of stars, of which only 10-15% were known variables at that time. Among high-amplitude variables, about 90-95% were already known, but the remaining 5-10% still represented a substantial number of objects. Especially when ZTF came into play, new cataclysmic variables could receieve a certain variability type than just "CV" or "UG". Some types (eg. UGER or NL/VY) increased the amount of known objects (based on VSX) by 25-100% at that time. The MGAB catalog contains mostly objects based on data mining, yet the variability of every object in it could be detected with the equipment mentioned above. My personal record for a stacked image is magnitude 22, and the shallowest positively detected exoplanet transit was about 0.005 mag.
A few years later, new catalogs changed everything. That initial 10-15% turned into 95%. There was almost nothing left to discover. Searching for 5-10 genuinely interesting variables was no longer worth the effort - especially when, in the past, the same amount of work yielded 50-100 objects. That time was not wasted, as those discoveries became part of the MGAB catalog. Not today, if doubled time spent would only increase by the amount by 10%.
For high-amplitude variables, probably only transients are left (or LSST/Vera Rubin faint variables, but unreachable for me using own telescope) and large survey observatories are racing to detect them. Even magnitude outliers, where as little as 0.01-0.1% of them may indicate rare eclipses or UV Ceti-type eruptions, can now be analyzed frame by frame using automated modern pipelines - the speed of modern technology allow this. The golden age of amateur discovery of new variable stars has ended. Today, discovery is driven almost entirely by automated surveys, where human impact on detection and classification is very limited or almost none.
Today, it is still possible to find very interesting behavior in already known variables - objects that were once placed into the “boring bag.” Am I interested in them? Honestly, no. I have a very strict rule: I do not analyze what is not mine. Others may approach this differently, and that is perfectly valid. This page exists partly to help myself organize follow-up of objects in the MGAB catalog.
It must take some time after a variable is published to say who the discoverer is. Why? No one should be expected to search the entire internet to verify whether an object was already known. A reasonable starting point is VizieR, which contains the largest accessible collection of variable star catalogs and includes regularly updated VSX data. If you enter the coordinates of an MGAB object today, it's been there for years now. Because VizieR also shows the VSX catalog.
So, if a paper published today does not mention my work, there are usually two possibilities: discovery credit was not prioritized, or the search was not deep enough... sometimes simply because the authors did not know where or how to look. Based on discussions with professional astronomers, including a conversation with Paula Szkody during an AAVSO webinar, the second explanation is entirely plausible. This was also one reason I pushed discoveries to my limits, trying to identify as many interesting variables as possible, knowing that some would eventually require inclusion in scientific papers. So many, to have a significant statistical base to say if amateur work is really seen in scientific astronomical community. As a few years passed while writting this, in how many papers there was a mention? About 40%. I am grateful to those authors - big thanks and respect!
However, VizieR alone is not sufficient. Variables can also be found in SIMBAD, and over time I began using it routinely. Then I discovered that the Transient Name Server (TNS) contained objects I had reported as mine, which required another review of discovery credit. There are likely additional databases I am still unaware of (and may never discover) sometimes simply due to language barriers.
Now, years after my last variable star discovery, I can finally close this subject. I no longer need to deal with the latest catalogs, projects, or AI-based classification pipelines that started recently into play. In short, as an amateur, I place very high value on discovery credit - probably more than I should. Time spent verifying credit could often be better spent on stellar analysis, but for me, this needed to be resolved first.
I still monitor arXiv to see whether variable stars I reported years ago have received detailed analysis. It is always rewarding to read such work. I also track whether a star is mentioned with its MGAB identifier and I collect those papers in a physical binder. It is a tangible way to feel that the work had some lasting impact, perhaps even for the rest of my life. This was the case with NL/VY variables: based on the VSX list, I reported as many as possible with the hope that one might erupt as a nova during my lifetime. Unexpectedly, MGAB-V207 did so the following year!
In variable stars world, not everyone treats discovery credit as a priority. Most professional astronomers focus on different metrics: who authored the paper, where it was published, how many citations it received, and whether it contributes to career advancement through positions or grants. I understand that well.
Then there is me - an amateur astronomer working professionally as a dentist. I do not have access to giant telescopes, cannot easily write a 30-page paper, and do not apply for grants. Yet I can still find stars whose variability is not listed in any catalog and report them in a way similar to large surveys, just on a scale an amateur can manage. You may include five variables in a 25-page paper, but you would never publish a 25,000-page paper with 5,000 variables. Instead, discoveries are compressed into catalogs. There are also valid ways to publish discoveries outside traditional journals. And the work spent on variable stars means very little for my day job (in fact, the only thing that binds dentistry and astronomy is black holes!). So why would an amateur even work in finding new variables? Most likely to discover something, report that "there's a new variable I found, maybe it's worth further follow up?".
There is a definition that captures discovery in astronomy particularly well...
"Discovery is not done by discovery. You find many different things. None of them are discovery until you understand what it is."
Discovery credit usually becomes a problem when you find out that someone else reported the same thing earlier - sometimes differing by only one important detail. But the deceptively simple question “when was it done?” hides many possible answers. Depending on the context, “done” may mean:
When the journal issue containing the paper was officially released?
When did your paper first appear publicly (e.g., on arXiv)?
When did the referee approve the manuscript?
When the paper was first submitted to a journal?
When the variability of a certain object was noticed for the first time?
When did a public notice appear stating that variability exists, even if follow-up analysis was still ongoing?
Each of these moments has a different date. Not all of them apply to every variable star. Different people also accept different moments as the “real” discovery date. Online variable star submissions skip many of these steps. That is useful, but it also raises an important question: How much must you know before you can claim a discovery? There is no single answer. Everyone sets this limit differently. My own limit is probably lower than that seen in professional astronomy. I do not feel obligated to perform radial-velocity observations to determine the mass of the secondary or the orbital eccentricity, for example. If brightness changes are real and confirmed, with data allowing to measure period or state the variability type, that is enough for me.
You may check catalogs and papers and honestly believe that a star is new. Later, you find out that someone else also noticed the same variability, but did not publish it yet. Some projects take years. Others take one day. You report the star first and, for a while, your report contains the most information available. Later, a detailed paper appears, and your contribution appears much smaller.
So, who is the discoverer? In this situation, independent discovery makes sense. This is applied for MGAB-V249, for example.
Professional astronomers usually have no problem with this. Parallel work is common. If someone else publishes similar results, it is mentioned briefly and life goes on. Papers, citations, and impact are more important than discovery credit. For an amateur astronomer, this is very different...
All my follow-up work is done only on stars I discovered myself. Because of that, discovery credit decides what I will work on in the future. Over time, many MGAB stars needed discovery credit corrections. This does not mean mistakes were made at the beginning. The more you search, the more databases you find. Some catalogs were not easy to find. Some did not exist yet. Others did not appear in early searches.
Also, not all catalogs are the same. There are sets of confirmed variables or candidates. Or, variability was based on the magnitude error, or the light curve shape. Vetted by a human or by an artifical intelligence (especially those "early AI" systems). Contains details like period, but might be correct or not. A binary with a known spectroscopic companion, but eclipses of the secondary I just found and weren't reported earlier. At some point even all your discoveries are questionable, thus for common sense this needs to be sorted. These are just a few examples showing why the word “discovery” does not have a simple meaning. There are dozens of situations that need to be investigated differently.
After a while, even your own discoveries start to raise questions. This is why the word “discovery” is not simple.
So, in which cases should the discovery credit be revised?
Variability was never really shown.
A star is not variable just because its catalog error is large. Variability must be seen — in images or light curves.
The star was already listed as a candidate, and the signal is real.
I accept these cases. Even if it was not fully confirmed, a real periodic signal matters. However, when the classification is clearly incorrect and the entire catalog repeats the same issue, an independent discovery credit is given, as is also done in VSX.
A clear earlier paper exists.
If variability was properly explained earlier, credit should be updated. I probably found it too late, but it still can be easily corrected.
Do you consider yourself as a discoverer of a certain variable star, but it is already in the MGAB Variable Star Catalog and already listed as a personal discovery?
Feel free to contact with me if you want this to be corrected (you can find my e-mail in About Me section). It's very likely that I simply missed your report/submission/paper/publication. If I see a date confirming that you have done it earlier and if I easily could have found it when I submitted the variable star (but I didn't know it existed, as it wasn't in my checklist), this can be easily adjusted. It's also possible that I had already noticed, but haven't submitted a revision yet (as those are done in batch). Not for everyone the main goal is to keep variability flags listed in VSX, VizieR or Simbad. If the source is hardly reachable, the chance of being missed increases.
What has been done with certain catalogs?
The search for variable stars started in 2017 by exploring the neighbourhood of stars hosting an exoplanet transit. More than a hundred of variables were submitted to the Variable Star Index individually using own photometry data combined with survey ones (eg. ASAS-SN Sky Patrol or SuperWASP). A few months later the search changed to data mining using survey data. It was more efficient - I
Variability flags marked as “true,” “yes,” or “1,” based solely on photometric errors, are not treated as discoveries, as none of these confirm that an object is indeed variable. All MGAB stars were checked manually by eye (even on single images, if needed) to make sure variables are indeed real.
Objects with a general classification based on spectroscopy or survey magnitudes are not treated as discoveries. Example: Young Stellar Object (YSO) from SIMBAD based only on IR-excess or H-alpha emission isn't enough to claim if the discovery, as photometric variability (in the paper giving such designation) was not mentioned. This case clearly shows that the MGAB project is about discovering variability in brightness, not object classification.
Variable stars from the ATLAS catalog (2018) are treated as full discoveries only when are not marked as 'dubious'. There is only one exception, MGAB-V266, a HW Virginis type binary with total eclipses. At the moment of discovery, second ever known (and shortly soon, MGAB-V283, third) such case. It was reported in the ATLAS catalog as "PULSE" type with doubled period, and does not confirm presence of eclipses. Because of uniqueness of the target, also presenting ETVs, the credit is given as an independent discovery. That's a significantly important object, where primary classification by ATLAS was very underestimated.
Gaia DR2 variables are all excluded even if their classification is incorrect (eg. DSCT_SXPHE for cataclysmic variables or MIRA for Young Stellar Objects). Further Gaia releases are not considered, as they are all too recent.
RR Lyrae candidates from Pan-STARRS1 are all treated as independent discoveries and both names will be always shown in the VSX (Sesar and Murawski). The main reason is, the catalog has a very high amount of incorrect classifications, which were often absurd, that could be easily be filtered during vetting. This applies especially to cataclysmic variables, where PANSTARRS-1 own data clearly show high amplitude variables with ranges impossible for RRAB and RRC stars. About 20% of CVs in the MGAB catalog are independent discoveries with this catalog, after cross-matching. My personal opinion is, it would be an exaggeration if I skipped all those really interesting stars for analysis and follow-up, only because of the quality of catalog (it's not the MGAB-V266 level as mentioned above, but any CV can result with interesting variability in the future), but leaving the independent discovery credit is plausible, as the variability signal is indeed real.
ZTF variable star candidates from the paper "A catalog of over ten million variable source candidates in ZTF data release 1" is another story. Thankfully, it contains a catalog of candidates, so definitely does not fit the scheme I previously written about. But it mentions that about 60 dwarf novae were identified. Based on submission dates, this could apply to MGAB variable stars between V3371 and V3774, so one could be revised. Unfortunately, the whole catalog was released with difficult access, without publishing to online sets like VizieR. And probably won't change, because the dataset compared to today ones, is obsolete. What’s more, the URL to download the catalog is no longer working, so the information required to adjust the discovery credit is simply erased.
The official ZTF lists (alerce.online & lasair.roe.ac.uk) will not be considered as discoveries unless they are published in other catalogs mentioned above. The discovery date given on these sites corresponds to the first ever positive detection of the target, not to a formal variability discovery report. Variables are not asteroids, where the credit scheme works that way. All MGAB variables were identified and submitted during 2019–2020, prior to the appearance of the ALeRCE broker publication and at a time when both ALeRCE and Lasair were still in a very early, exploratory stage of development and were not issuing formal discovery reports to the Transient Name Server (TNS). This alone is sufficient justification for not treating broker alert lists from that period as discovery claims. The TNS, that worked separately, already existed at that time, and a small fraction of MGAB objects (of order 2% of CVs) were later reported to TNS by ZTF. In such cases (e.g. MGAB-V301), the credited discoverer was corrected according to the accepted TNS discovery report, even though the object had been detected earlier and, in some instances, listed in other catalogs such as VSX.
Not all authors of papers submit their findings to SIMBAD, VizieR or even VSX. If a variable star with a reasonable variability explaination in a previously published paper will be found after the submission with MGAB as a main name, it will be corrected.
In VSX, a proper credit from using datapoints coming from a certain survey in plots is given. In this website, they can be found below each light curves page.
Discovery dates for all objects:
MGAB-V1 to MGAB-V300 - all these objects were submitted manually to the VSX, thus each date is different for each object. Identifications are not in chronological order, especially above V200. The earliest date is for MGAB-V1 (the first discovery, 15.08.2017) and the soonest for MGAB-V209 and MGAB-V210 (both 07.02.2020). The same thing is applied to objects from the paper "The discovery of eleven nova-like VY Scl type stars" (Murawski, G.; 2019). Due to delays in publication, I decided to submit them to the VSX earlier, than the paper publication date.
MGAB-V301 to MGAB-V550 - 25.09.2019
MGAB-V551 to MGAB-V1120 - 15.10.2019
MGAB-V1121 to MGAB-V1146 - 11.01.2020
MGAB-V1147 to MGAB-V1290 - 27.01.2020
MGAB-V1291 to MGAB-V3361 - 21.02.2020
MGAB-V3362 to MGAB-V3370 - 15.05.2020
MGAB-V3371 to MGAB-V3774 - 20.12.2020 (except MGAB-V3390, MGAB-V3395, MGAB-V3396 - individual submissions to the VSX)
MGAB-V3775 - 15.02.2019 (old object with designation changed due to MGAB-V216 being submitted twice)
MGAB-V3776 - 15.04.2019 (old object with designation changed due to MGAB-V223 being submitted twice)
MGAB-V3775 to MGAB-V3813 - 14.03.2022